Automatic merge of rsync://rsync.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git branch HEAD

80 files changed, 1433 insertions, 1135 deletions

diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c
index 64e450dddb4..167fd89f870 100644
--- a/arch/alpha/kernel/osf_sys.c
+++ b/arch/alpha/kernel/osf_sys.c
@@ -1150,16 +1150,13 @@ osf_usleep_thread(struct timeval32 __user *sleep, struct timeval32 __user *remai
 	if (get_tv32(&tmp, sleep))
 		goto fault;
 
-	ticks = tmp.tv_usec;
-	ticks = (ticks + (1000000 / HZ) - 1) / (1000000 / HZ);
-	ticks += tmp.tv_sec * HZ;
+	ticks = timeval_to_jiffies(&tmp);
 
 	current->state = TASK_INTERRUPTIBLE;
 	ticks = schedule_timeout(ticks);
 
 	if (remain) {
-		tmp.tv_sec = ticks / HZ;
-		tmp.tv_usec = ticks % HZ;
+		jiffies_to_timeval(ticks, &tmp);
 		if (put_tv32(remain, &tmp))
 			goto fault;
 	}
diff --git a/arch/arm/mach-s3c2410/clock.c b/arch/arm/mach-s3c2410/clock.c
index e23f534d4e1..8d986b8401c 100644
--- a/arch/arm/mach-s3c2410/clock.c
+++ b/arch/arm/mach-s3c2410/clock.c
@@ -478,7 +478,7 @@ static int s3c2440_clk_add(struct sys_device *sysdev)
 {
 	unsigned long upllcon = __raw_readl(S3C2410_UPLLCON);
 
-	s3c2440_clk_upll.rate = s3c2410_get_pll(upllcon, clk_xtal.rate) * 2;
+	s3c2440_clk_upll.rate = s3c2410_get_pll(upllcon, clk_xtal.rate);
 
 	printk("S3C2440: Clock Support, UPLL %ld.%03ld MHz\n",
 	       print_mhz(s3c2440_clk_upll.rate));
diff --git a/arch/arm/mach-s3c2410/s3c2440.c b/arch/arm/mach-s3c2410/s3c2440.c
index 9a8cc5ae225..d4c8281b55f 100644
--- a/arch/arm/mach-s3c2410/s3c2440.c
+++ b/arch/arm/mach-s3c2410/s3c2440.c
@@ -192,9 +192,11 @@ void __init s3c2440_map_io(struct map_desc *mach_desc, int size)
 
 	iotable_init(s3c2440_iodesc, ARRAY_SIZE(s3c2440_iodesc));
 	iotable_init(mach_desc, size);
+
 	/* rename any peripherals used differing from the s3c2410 */
 
-	s3c_device_i2c.name = "s3c2440-i2c";
+	s3c_device_i2c.name  = "s3c2440-i2c";
+	s3c_device_nand.name = "s3c2440-nand";
 
 	/* change irq for watchdog */
 
@@ -225,7 +227,7 @@ void __init s3c2440_init_clocks(int xtal)
 		break;
 
 	case S3C2440_CLKDIVN_HDIVN_2:
-		hdiv = 1;
+		hdiv = 2;
 		break;
 
 	case S3C2440_CLKDIVN_HDIVN_4_8:
diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig
index c4fc6be629d..48bac7da8c7 100644
--- a/arch/arm/mm/Kconfig
+++ b/arch/arm/mm/Kconfig
@@ -412,21 +412,20 @@ config CPU_BPREDICT_DISABLE
 
 config TLS_REG_EMUL
 	bool
-	default y if (SMP || CPU_32v6) && (CPU_32v5 || CPU_32v4 || CPU_32v3)
+	default y if SMP && (CPU_32v5 || CPU_32v4 || CPU_32v3)
 	help
-	  We might be running on an ARMv6+ processor which should have the TLS
-	  register but for some reason we can't use it, or maybe an SMP system
-	  using a pre-ARMv6 processor (there are apparently a few prototypes
-	  like that in existence) and therefore access to that register must
-	  be emulated.
+	  An SMP system using a pre-ARMv6 processor (there are apparently
+	  a few prototypes like that in existence) and therefore access to
+	  that required register must be emulated.
 
 config HAS_TLS_REG
 	bool
-	depends on CPU_32v6
-	default y if !TLS_REG_EMUL
+	depends on !TLS_REG_EMUL
+	default y if SMP || CPU_32v7
 	help
 	  This selects support for the CP15 thread register.
-	  It is defined to be available on ARMv6 or later.  If a particular
-	  ARMv6 or later CPU doesn't support it then it must omc;ide "select
-	  TLS_REG_EMUL" along with its other caracteristics.
+	  It is defined to be available on some ARMv6 processors (including
+	  all SMP capable ARMv6's) or later processors.  User space may
+	  assume directly accessing that register and always obtain the
+	  expected value only on ARMv7 and above.
 
diff --git a/arch/arm/mm/copypage-v4mc.S b/arch/arm/mm/copypage-v4mc.S
deleted file mode 100644
index 305af3dab3d..00000000000
--- a/arch/arm/mm/copypage-v4mc.S
+++ /dev/null
@@ -1,80 +0,0 @@
-/*
- *  linux/arch/arm/lib/copy_page-armv4mc.S
- *
- *  Copyright (C) 1995-2001 Russell King
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- *
- *  ASM optimised string functions
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/constants.h>
-
-	.text
-	.align	5
-/*
- * ARMv4 mini-dcache optimised copy_user_page
- *
- * We flush the destination cache lines just before we write the data into the
- * corresponding address.  Since the Dcache is read-allocate, this removes the
- * Dcache aliasing issue.  The writes will be forwarded to the write buffer,
- * and merged as appropriate.
- *
- * Note: We rely on all ARMv4 processors implementing the "invalidate D line"
- * instruction.  If your processor does not supply this, you have to write your
- * own copy_user_page that does the right thing.
- */
-ENTRY(v4_mc_copy_user_page)
-	stmfd	sp!, {r4, lr}			@ 2
-	mov	r4, r0
-	mov	r0, r1
-	bl	map_page_minicache
-	mov	r1, #PAGE_SZ/64			@ 1
-	ldmia	r0!, {r2, r3, ip, lr}		@ 4
-1:	mcr	p15, 0, r4, c7, c6, 1		@ 1   invalidate D line
-	stmia	r4!, {r2, r3, ip, lr}		@ 4
-	ldmia	r0!, {r2, r3, ip, lr}		@ 4+1
-	stmia	r4!, {r2, r3, ip, lr}		@ 4
-	ldmia	r0!, {r2, r3, ip, lr}		@ 4
-	mcr	p15, 0, r4, c7, c6, 1		@ 1   invalidate D line
-	stmia	r4!, {r2, r3, ip, lr}		@ 4
-	ldmia	r0!, {r2, r3, ip, lr}		@ 4
-	subs	r1, r1, #1			@ 1
-	stmia	r4!, {r2, r3, ip, lr}		@ 4
-	ldmneia	r0!, {r2, r3, ip, lr}		@ 4
-	bne	1b				@ 1
-	ldmfd	sp!, {r4, pc}			@ 3
-
-	.align	5
-/*
- * ARMv4 optimised clear_user_page
- *
- * Same story as above.
- */
-ENTRY(v4_mc_clear_user_page)
-	str	lr, [sp, #-4]!
-	mov	r1, #PAGE_SZ/64			@ 1
-	mov	r2, #0				@ 1
-	mov	r3, #0				@ 1
-	mov	ip, #0				@ 1
-	mov	lr, #0				@ 1
-1:	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	subs	r1, r1, #1			@ 1
-	bne	1b				@ 1
-	ldr	pc, [sp], #4
-
-	__INITDATA
-
-	.type	v4_mc_user_fns, #object
-ENTRY(v4_mc_user_fns)
-	.long	v4_mc_clear_user_page
-	.long	v4_mc_copy_user_page
-	.size	v4_mc_user_fns, . - v4_mc_user_fns
diff --git a/arch/arm/mm/copypage-v4mc.c b/arch/arm/mm/copypage-v4mc.c
new file mode 100644
index 00000000000..fc69dccdace
--- /dev/null
+++ b/arch/arm/mm/copypage-v4mc.c
@@ -0,0 +1,111 @@
+/*
+ *  linux/arch/arm/lib/copypage-armv4mc.S
+ *
+ *  Copyright (C) 1995-2005 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This handles the mini data cache, as found on SA11x0 and XScale
+ * processors.  When we copy a user page page, we map it in such a way
+ * that accesses to this page will not touch the main data cache, but
+ * will be cached in the mini data cache.  This prevents us thrashing
+ * the main data cache on page faults.
+ */
+#include <linux/init.h>
+#include <linux/mm.h>
+
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/tlbflush.h>
+
+/*
+ * 0xffff8000 to 0xffffffff is reserved for any ARM architecture
+ * specific hacks for copying pages efficiently.
+ */
+#define minicache_pgprot __pgprot(L_PTE_PRESENT | L_PTE_YOUNG | \
+				  L_PTE_CACHEABLE)
+
+#define TOP_PTE(x)	pte_offset_kernel(top_pmd, x)
+
+static DEFINE_SPINLOCK(minicache_lock);
+
+/*
+ * ARMv4 mini-dcache optimised copy_user_page
+ *
+ * We flush the destination cache lines just before we write the data into the
+ * corresponding address.  Since the Dcache is read-allocate, this removes the
+ * Dcache aliasing issue.  The writes will be forwarded to the write buffer,
+ * and merged as appropriate.
+ *
+ * Note: We rely on all ARMv4 processors implementing the "invalidate D line"
+ * instruction.  If your processor does not supply this, you have to write your
+ * own copy_user_page that does the right thing.
+ */
+static void __attribute__((naked))
+mc_copy_user_page(void *from, void *to)
+{
+	asm volatile(
+	"stmfd	sp!, {r4, lr}			@ 2\n\
+	mov	r4, %2				@ 1\n\
+	ldmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+1:	mcr	p15, 0, %1, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	%1!, {r2, r3, ip, lr}		@ 4\n\
+	ldmia	%0!, {r2, r3, ip, lr}		@ 4+1\n\
+	stmia	%1!, {r2, r3, ip, lr}		@ 4\n\
+	ldmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	mcr	p15, 0, %1, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	%1!, {r2, r3, ip, lr}		@ 4\n\
+	ldmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	subs	r4, r4, #1			@ 1\n\
+	stmia	%1!, {r2, r3, ip, lr}		@ 4\n\
+	ldmneia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	bne	1b				@ 1\n\
+	ldmfd	sp!, {r4, pc}			@ 3"
+	:
+	: "r" (from), "r" (to), "I" (PAGE_SIZE / 64));
+}
+
+void v4_mc_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	spin_lock(&minicache_lock);
+
+	set_pte(TOP_PTE(0xffff8000), pfn_pte(__pa(kfrom) >> PAGE_SHIFT, minicache_pgprot));
+	flush_tlb_kernel_page(0xffff8000);
+
+	mc_copy_user_page((void *)0xffff8000, kto);
+
+	spin_unlock(&minicache_lock);
+}
+
+/*
+ * ARMv4 optimised clear_user_page
+ */
+void __attribute__((naked))
+v4_mc_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm volatile(
+	"str	lr, [sp, #-4]!\n\
+	mov	r1, %0				@ 1\n\
+	mov	r2, #0				@ 1\n\
+	mov	r3, #0				@ 1\n\
+	mov	ip, #0				@ 1\n\
+	mov	lr, #0				@ 1\n\
+1:	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	subs	r1, r1, #1			@ 1\n\
+	bne	1b				@ 1\n\
+	ldr	pc, [sp], #4"
+	:
+	: "I" (PAGE_SIZE / 64));
+}
+
+struct cpu_user_fns v4_mc_user_fns __initdata = {
+	.cpu_clear_user_page	= v4_mc_clear_user_page, 
+	.cpu_copy_user_page	= v4_mc_copy_user_page,
+};
diff --git a/arch/arm/mm/copypage-v6.c b/arch/arm/mm/copypage-v6.c
index 694ac820885..a8c00236bd3 100644
--- a/arch/arm/mm/copypage-v6.c
+++ b/arch/arm/mm/copypage-v6.c
@@ -26,8 +26,8 @@
 #define to_address	(0xffffc000)
 #define to_pgprot	PAGE_KERNEL
 
-static pte_t *from_pte;
-static pte_t *to_pte;
+#define TOP_PTE(x)	pte_offset_kernel(top_pmd, x)
+
 static DEFINE_SPINLOCK(v6_lock);
 
 #define DCACHE_COLOUR(vaddr) ((vaddr & (SHMLBA - 1)) >> PAGE_SHIFT)
@@ -74,8 +74,8 @@ void v6_copy_user_page_aliasing(void *kto, const void *kfrom, unsigned long vadd
 	 */
 	spin_lock(&v6_lock);
 
-	set_pte(from_pte + offset, pfn_pte(__pa(kfrom) >> PAGE_SHIFT, from_pgprot));
-	set_pte(to_pte + offset, pfn_pte(__pa(kto) >> PAGE_SHIFT, to_pgprot));
+	set_pte(TOP_PTE(from_address) + offset, pfn_pte(__pa(kfrom) >> PAGE_SHIFT, from_pgprot));
+	set_pte(TOP_PTE(to_address) + offset, pfn_pte(__pa(kto) >> PAGE_SHIFT, to_pgprot));
 
 	from = from_address + (offset << PAGE_SHIFT);
 	to   = to_address + (offset << PAGE_SHIFT);
@@ -114,7 +114,7 @@ void v6_clear_user_page_aliasing(void *kaddr, unsigned long vaddr)
 	 */
 	spin_lock(&v6_lock);
 
-	set_pte(to_pte + offset, pfn_pte(__pa(kaddr) >> PAGE_SHIFT, to_pgprot));
+	set_pte(TOP_PTE(to_address) + offset, pfn_pte(__pa(kaddr) >> PAGE_SHIFT, to_pgprot));
 	flush_tlb_kernel_page(to);
 	clear_page((void *)to);
 
@@ -129,21 +129,6 @@ struct cpu_user_fns v6_user_fns __initdata = {
 static int __init v6_userpage_init(void)
 {
 	if (cache_is_vipt_aliasing()) {
-		pgd_t *pgd;
-		pmd_t *pmd;
-
-		pgd = pgd_offset_k(from_address);
-		pmd = pmd_alloc(&init_mm, pgd, from_address);
-		if (!pmd)
-			BUG();
-		from_pte = pte_alloc_kernel(&init_mm, pmd, from_address);
-		if (!from_pte)
-			BUG();
-
-		to_pte = pte_alloc_kernel(&init_mm, pmd, to_address);
-		if (!to_pte)
-			BUG();
-
 		cpu_user.cpu_clear_user_page = v6_clear_user_page_aliasing;
 		cpu_user.cpu_copy_user_page = v6_copy_user_page_aliasing;
 	}
@@ -151,5 +136,4 @@ static int __init v6_userpage_init(void)
 	return 0;
 }
 
-__initcall(v6_userpage_init);
-
+core_initcall(v6_userpage_init);
diff --git a/arch/arm/mm/flush.c b/arch/arm/mm/flush.c
index c6de48d8950..4085ed983e4 100644
--- a/arch/arm/mm/flush.c
+++ b/arch/arm/mm/flush.c
@@ -13,6 +13,29 @@
 
 #include <asm/cacheflush.h>
 #include <asm/system.h>
+#include <asm/tlbflush.h>
+
+#ifdef CONFIG_CPU_CACHE_VIPT
+#define ALIAS_FLUSH_START	0xffff4000
+
+#define TOP_PTE(x)	pte_offset_kernel(top_pmd, x)
+
+static void flush_pfn_alias(unsigned long pfn, unsigned long vaddr)
+{
+	unsigned long to = ALIAS_FLUSH_START + (CACHE_COLOUR(vaddr) << PAGE_SHIFT);
+
+	set_pte(TOP_PTE(to), pfn_pte(pfn, PAGE_KERNEL));
+	flush_tlb_kernel_page(to);
+
+	asm(	"mcrr	p15, 0, %1, %0, c14\n"
+	"	mcrr	p15, 0, %1, %0, c5\n"
+	    :
+	    : "r" (to), "r" (to + PAGE_SIZE - L1_CACHE_BYTES)
+	    : "cc");
+}
+#else
+#define flush_pfn_alias(pfn,vaddr)	do { } while (0)
+#endif
 
 static void __flush_dcache_page(struct address_space *mapping, struct page *page)
 {
@@ -37,6 +60,18 @@ static void __flush_dcache_page(struct address_space *mapping, struct page *page
 		return;
 
 	/*
+	 * This is a page cache page.  If we have a VIPT cache, we
+	 * only need to do one flush - which would be at the relevant
+	 * userspace colour, which is congruent with page->index.
+	 */
+	if (cache_is_vipt()) {
+		if (cache_is_vipt_aliasing())
+			flush_pfn_alias(page_to_pfn(page),
+					page->index << PAGE_CACHE_SHIFT);
+		return;
+	}
+
+	/*
 	 * There are possible user space mappings of this page:
 	 * - VIVT cache: we need to also write back and invalidate all user
 	 *   data in the current VM view associated with this page.
@@ -57,8 +92,6 @@ static void __flush_dcache_page(struct address_space *mapping, struct page *page
 			continue;
 		offset = (pgoff - mpnt->vm_pgoff) << PAGE_SHIFT;
 		flush_cache_page(mpnt, mpnt->vm_start + offset, page_to_pfn(page));
-		if (cache_is_vipt())
-			break;
 	}
 	flush_dcache_mmap_unlock(mapping);
 }
diff --git a/arch/arm/mm/mm-armv.c b/arch/arm/mm/mm-armv.c
index 585dfb8e20b..2c2b93d77d4 100644
--- a/arch/arm/mm/mm-armv.c
+++ b/arch/arm/mm/mm-armv.c
@@ -37,6 +37,8 @@ pgprot_t pgprot_kernel;
 
 EXPORT_SYMBOL(pgprot_kernel);
 
+pmd_t *top_pmd;
+
 struct cachepolicy {
 	const char	policy[16];
 	unsigned int	cr_mask;
@@ -142,6 +144,16 @@ __setup("noalign", noalign_setup);
 
 #define FIRST_KERNEL_PGD_NR	(FIRST_USER_PGD_NR + USER_PTRS_PER_PGD)
 
+static inline pmd_t *pmd_off(pgd_t *pgd, unsigned long virt)
+{
+	return pmd_offset(pgd, virt);
+}
+
+static inline pmd_t *pmd_off_k(unsigned long virt)
+{
+	return pmd_off(pgd_offset_k(virt), virt);
+}
+
 /*
  * need to get a 16k page for level 1
  */
@@ -220,7 +232,7 @@ void free_pgd_slow(pgd_t *pgd)
 		return;
 
 	/* pgd is always present and good */
-	pmd = (pmd_t *)pgd;
+	pmd = pmd_off(pgd, 0);
 	if (pmd_none(*pmd))
 		goto free;
 	if (pmd_bad(*pmd)) {
@@ -246,9 +258,8 @@ free:
 static inline void
 alloc_init_section(unsigned long virt, unsigned long phys, int prot)
 {
-	pmd_t *pmdp;
+	pmd_t *pmdp = pmd_off_k(virt);
 
-	pmdp = pmd_offset(pgd_offset_k(virt), virt);
 	if (virt & (1 << 20))
 		pmdp++;
 
@@ -283,11 +294,9 @@ alloc_init_supersection(unsigned long virt, unsigned long phys, int prot)
 static inline void
 alloc_init_page(unsigned long virt, unsigned long phys, unsigned int prot_l1, pgprot_t prot)
 {
-	pmd_t *pmdp;
+	pmd_t *pmdp = pmd_off_k(virt);
 	pte_t *ptep;
 
-	pmdp = pmd_offset(pgd_offset_k(virt), virt);
-
 	if (pmd_none(*pmdp)) {
 		unsigned long pmdval;
 		ptep = alloc_bootmem_low_pages(2 * PTRS_PER_PTE *
@@ -310,7 +319,7 @@ alloc_init_page(unsigned long virt, unsigned long phys, unsigned int prot_l1, pg
  */
 static inline void clear_mapping(unsigned long virt)
 {
-	pmd_clear(pmd_offset(pgd_offset_k(virt), virt));
+	pmd_clear(pmd_off_k(virt));
 }
 
 struct mem_types {
@@ -578,7 +587,7 @@ void setup_mm_for_reboot(char mode)
 			 PMD_TYPE_SECT;
 		if (cpu_arch <= CPU_ARCH_ARMv5)
 			pmdval |= PMD_BIT4;
-		pmd = pmd_offset(pgd + i, i << PGDIR_SHIFT);
+		pmd = pmd_off(pgd, i << PGDIR_SHIFT);
 		pmd[0] = __pmd(pmdval);
 		pmd[1] = __pmd(pmdval + (1 << (PGDIR_SHIFT - 1)));
 		flush_pmd_entry(pmd);
@@ -675,6 +684,8 @@ void __init memtable_init(struct meminfo *mi)
 
 	flush_cache_all();
 	flush_tlb_all();
+
+	top_pmd = pmd_off_k(0xffff0000);
 }
 
 /*
diff --git a/arch/i386/kernel/cpu/amd.c b/arch/i386/kernel/cpu/amd.c
index 16dbc4151be..73aeaf5a9d4 100644
--- a/arch/i386/kernel/cpu/amd.c
+++ b/arch/i386/kernel/cpu/amd.c
@@ -24,9 +24,6 @@ __asm__(".align 4\nvide: ret");
 
 static void __init init_amd(struct cpuinfo_x86 *c)
 {
-#ifdef CONFIG_X86_SMP
-	int cpu = c == &boot_cpu_data ? 0 : c - cpu_data;
-#endif
 	u32 l, h;
 	int mbytes = num_physpages >> (20-PAGE_SHIFT);
 	int r;
@@ -198,14 +195,19 @@ static void __init init_amd(struct cpuinfo_x86 *c)
 			c->x86_num_cores = 1;
 	}
 
-#ifdef CONFIG_X86_SMP
+#ifdef CONFIG_X86_HT
 	/*
 	 * On a AMD dual core setup the lower bits of the APIC id
 	 * distingush the cores.  Assumes number of cores is a power
 	 * of two.
 	 */
 	if (c->x86_num_cores > 1) {
-		cpu_core_id[cpu] = cpu >> hweight32(c->x86_num_cores - 1);
+		int cpu = smp_processor_id();
+		unsigned bits = 0;
+		while ((1 << bits) < c->x86_num_cores)
+			bits++;
+		cpu_core_id[cpu] = phys_proc_id[cpu] & ((1<<bits)-1);
+		phys_proc_id[cpu] >>= bits;
 		printk(KERN_INFO "CPU %d(%d) -> Core %d\n",
 		       cpu, c->x86_num_cores, cpu_core_id[cpu]);
 	}
diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c
index 6be0310e3cd..d199e525680 100644
--- a/arch/i386/kernel/cpu/common.c
+++ b/arch/i386/kernel/cpu/common.c
@@ -243,6 +243,10 @@ static void __init early_cpu_detect(void)
 	}
 
 	early_intel_workaround(c);
+
+#ifdef CONFIG_X86_HT
+	phys_proc_id[smp_processor_id()] = (cpuid_ebx(1) >> 24) & 0xff;
+#endif
 }
 
 void __init generic_identify(struct cpuinfo_x86 * c)
diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c
index cbea7ac582e..35bfe138cb1 100644
--- a/arch/i386/kernel/smpboot.c
+++ b/arch/i386/kernel/smpboot.c
@@ -888,6 +888,7 @@ void *xquad_portio;
 
 cpumask_t cpu_sibling_map[NR_CPUS] __cacheline_aligned;
 cpumask_t cpu_core_map[NR_CPUS] __cacheline_aligned;
+EXPORT_SYMBOL(cpu_core_map);
 
 static void __init smp_boot_cpus(unsigned int max_cpus)
 {
diff --git a/arch/i386/mach-voyager/voyager_smp.c b/arch/i386/mach-voyager/voyager_smp.c
index 903d739ca74..a6e0ddd65bd 100644
--- a/arch/i386/mach-voyager/voyager_smp.c
+++ b/arch/i386/mach-voyager/voyager_smp.c
@@ -97,7 +97,6 @@ static void ack_vic_irq(unsigned int irq);
 static void vic_enable_cpi(void);
 static void do_boot_cpu(__u8 cpuid);
 static void do_quad_bootstrap(void);
-static inline void wrapper_smp_local_timer_interrupt(struct pt_regs *);
 
 int hard_smp_processor_id(void);
 
@@ -126,6 +125,14 @@ send_QIC_CPI(__u32 cpuset, __u8 cpi)
 }
 
 static inline void
+wrapper_smp_local_timer_interrupt(struct pt_regs *regs)
+{
+	irq_enter();
+	smp_local_timer_interrupt(regs);
+	irq_exit();
+}
+
+static inline void
 send_one_CPI(__u8 cpu, __u8 cpi)
 {
 	if(voyager_quad_processors & (1<<cpu))
@@ -1249,14 +1256,6 @@ smp_vic_timer_interrupt(struct pt_regs *regs)
 	smp_local_timer_interrupt(regs);
 }
 
-static inline void
-wrapper_smp_local_timer_interrupt(struct pt_regs *regs)
-{
-	irq_enter();
-	smp_local_timer_interrupt(regs);
-	irq_exit();
-}
-
 /* local (per CPU) timer interrupt.  It does both profiling and
  * process statistics/rescheduling.
  *
diff --git a/arch/i386/mm/ioremap.c b/arch/i386/mm/ioremap.c
index db06f739991..ab542792b27 100644
--- a/arch/i386/mm/ioremap.c
+++ b/arch/i386/mm/ioremap.c
@@ -238,19 +238,21 @@ void iounmap(volatile void __iomem *addr)
 			addr < phys_to_virt(ISA_END_ADDRESS))
 		return;
 
-	p = remove_vm_area((void *) (PAGE_MASK & (unsigned long __force) addr));
+	write_lock(&vmlist_lock);
+	p = __remove_vm_area((void *) (PAGE_MASK & (unsigned long __force) addr));
 	if (!p) { 
-		printk("__iounmap: bad address %p\n", addr);
-		return;
+		printk("iounmap: bad address %p\n", addr);
+		goto out_unlock;
 	}
 
 	if ((p->flags >> 20) && p->phys_addr < virt_to_phys(high_memory) - 1) {
-		/* p->size includes the guard page, but cpa doesn't like that */
 		change_page_attr(virt_to_page(__va(p->phys_addr)),
 				 p->size >> PAGE_SHIFT,
 				 PAGE_KERNEL);
 		global_flush_tlb();
 	} 
+out_unlock:
+	write_unlock(&vmlist_lock);
 	kfree(p); 
 }
 
diff --git a/arch/i386/pci/fixup.c b/arch/i386/pci/fixup.c
index be52c5ac4e0..8e8e895e1b5 100644
--- a/arch/i386/pci/fixup.c
+++ b/arch/i386/pci/fixup.c
@@ -253,7 +253,7 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE2, pci
 #define MAX_PCIEROOT	6
 static int quirk_aspm_offset[MAX_PCIEROOT << 3];
 
-#define GET_INDEX(a, b) (((a - PCI_DEVICE_ID_INTEL_MCH_PA) << 3) + b)
+#define GET_INDEX(a, b) ((((a) - PCI_DEVICE_ID_INTEL_MCH_PA) << 3) + ((b) & 7))
 
 static int quirk_pcie_aspm_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *value)
 {
diff --git a/arch/ia64/ia32/ia32_ioctl.c b/arch/ia64/ia32/ia32_ioctl.c
index 9845dabe261..164b211f417 100644
--- a/arch/ia64/ia32/ia32_ioctl.c
+++ b/arch/ia64/ia32/ia32_ioctl.c
@@ -13,7 +13,6 @@
   
 #define	INCLUDES
 #include "compat_ioctl.c"
-#include <asm/ioctl32.h>
 
 #define IOCTL_NR(a)	((a) & ~(_IOC_SIZEMASK << _IOC_SIZESHIFT))
 
diff --git a/arch/mips/vr41xx/common/pmu.c b/arch/mips/vr41xx/common/pmu.c
index c5f1043de93..53166f3598b 100644
--- a/arch/mips/vr41xx/common/pmu.c
+++ b/arch/mips/vr41xx/common/pmu.c
@@ -1,7 +1,7 @@
 /*
  *  pmu.c, Power Management Unit routines for NEC VR4100 series.
  *
- *  Copyright (C) 2003-2004  Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
+ *  Copyright (C) 2003-2005  Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
  *
  *  This program is free software; you can redistribute it and/or modify
  *  it under the terms of the GNU General Public License as published by
@@ -17,7 +17,9 @@
  *  along with this program; if not, write to the Free Software
  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
+#include <linux/errno.h>
 #include <linux/init.h>
+#include <linux/ioport.h>
 #include <linux/kernel.h>
 #include <linux/smp.h>
 #include <linux/types.h>
@@ -27,20 +29,31 @@
 #include <asm/reboot.h>
 #include <asm/system.h>
 
-#define PMUCNT2REG	KSEG1ADDR(0x0f0000c6)
+#define PMU_TYPE1_BASE	0x0b0000a0UL
+#define PMU_TYPE1_SIZE	0x0eUL
+
+#define PMU_TYPE2_BASE	0x0f0000c0UL
+#define PMU_TYPE2_SIZE	0x10UL
+
+#define PMUCNT2REG	0x06
  #define SOFTRST	0x0010
 
+static void __iomem *pmu_base;
+
+#define pmu_read(offset)		readw(pmu_base + (offset))
+#define pmu_write(offset, value)	writew((value), pmu_base + (offset))
+
 static inline void software_reset(void)
 {
-	uint16_t val;
+	uint16_t pmucnt2;
 
 	switch (current_cpu_data.cputype) {
 	case CPU_VR4122:
 	case CPU_VR4131:
 	case CPU_VR4133:
-		val = readw(PMUCNT2REG);
-		val |= SOFTRST;
-		writew(val, PMUCNT2REG);
+		pmucnt2 = pmu_read(PMUCNT2REG);
+		pmucnt2 |= SOFTRST;
+		pmu_write(PMUCNT2REG, pmucnt2);
 		break;
 	default:
 		break;
@@ -71,6 +84,34 @@ static void vr41xx_power_off(void)
 
 static int __init vr41xx_pmu_init(void)
 {
+	unsigned long start, size;
+
+	switch (current_cpu_data.cputype) {
+	case CPU_VR4111:
+	case CPU_VR4121:
+		start = PMU_TYPE1_BASE;
+		size = PMU_TYPE1_SIZE;
+		break;
+	case CPU_VR4122:
+	case CPU_VR4131:
+	case CPU_VR4133:
+		start = PMU_TYPE2_BASE;
+		size = PMU_TYPE2_SIZE;
+		break;
+	default:
+		printk("Unexpected CPU of NEC VR4100 series\n");
+		return -ENODEV;
+	}
+
+	if (request_mem_region(start, size, "PMU") == NULL)
+		return -EBUSY;
+
+	pmu_base = ioremap(start, size);
+	if (pmu_base == NULL) {
+		release_mem_region(start, size);
+		return -EBUSY;
+	}
+
 	_machine_restart = vr41xx_restart;
 	_machine_halt = vr41xx_halt;
 	_machine_power_off = vr41xx_power_off;
@@ -78,4 +119,4 @@ static int __init vr41xx_pmu_init(void)
 	return 0;
 }
 
-early_initcall(vr41xx_pmu_init);
+core_initcall(vr41xx_pmu_init);
diff --git a/arch/ppc/kernel/head_44x.S b/arch/ppc/kernel/head_44x.S
index 9b6a8e51365..6c7ae605246 100644
--- a/arch/ppc/kernel/head_44x.S
+++ b/arch/ppc/kernel/head_44x.S
@@ -330,8 +330,9 @@ interrupt_base:
 	/* If we are faulting a kernel address, we have to use the
 	 * kernel page tables.
 	 */
-	andis.	r11, r10, 0x8000
-	beq	3f
+	lis	r11, TASK_SIZE@h
+	cmplw	r10, r11
+	blt+	3f
 	lis	r11, swapper_pg_dir@h
 	ori	r11, r11, swapper_pg_dir@l
 
@@ -464,8 +465,9 @@ interrupt_base:
 	/* If we are faulting a kernel address, we have to use the
 	 * kernel page tables.
 	 */
-	andis.	r11, r10, 0x8000
-	beq	3f
+	lis	r11, TASK_SIZE@h
+	cmplw	r10, r11
+	blt+	3f
 	lis	r11, swapper_pg_dir@h
 	ori	r11, r11, swapper_pg_dir@l
 
@@ -533,8 +535,9 @@ interrupt_base:
 	/* If we are faulting a kernel address, we have to use the
 	 * kernel page tables.
 	 */
-	andis.	r11, r10, 0x8000
-	beq	3f
+	lis	r11, TASK_SIZE@h
+	cmplw	r10, r11
+	blt+	3f
 	lis	r11, swapper_pg_dir@h
 	ori	r11, r11, swapper_pg_dir@l
 
diff --git a/arch/ppc/kernel/setup.c b/arch/ppc/kernel/setup.c
index 5dfb42f1a15..5c20266e3b1 100644
--- a/arch/ppc/kernel/setup.c
+++ b/arch/ppc/kernel/setup.c
@@ -499,7 +499,7 @@ static int __init set_preferred_console(void)
 {
 	struct device_node *prom_stdout;
 	char *name;
-	int offset;
+	int offset = 0;
 
 	if (of_stdout_device == NULL)
 		return -ENODEV;
@@ -753,6 +753,8 @@ void __init setup_arch(char **cmdline_p)
 	strlcpy(saved_command_line, cmd_line, COMMAND_LINE_SIZE);
 	*cmdline_p = cmd_line;
 
+	parse_early_param();
+
 	/* set up the bootmem stuff with available memory */
 	do_init_bootmem();
 	if ( ppc_md.progress ) ppc_md.progress("setup_arch: bootmem", 0x3eab);
diff --git a/arch/ppc/lib/string.S b/arch/ppc/lib/string.S
index 8d08a2eb225..36c9b97fd92 100644
--- a/arch/ppc/lib/string.S
+++ b/arch/ppc/lib/string.S
@@ -446,6 +446,7 @@ _GLOBAL(__copy_tofrom_user)
 #ifdef CONFIG_8xx
 	/* Don't use prefetch on 8xx */
 	mtctr	r0
+	li	r0,0
 53:	COPY_16_BYTES_WITHEX(0)
 	bdnz	53b
 
@@ -564,7 +565,9 @@ _GLOBAL(__copy_tofrom_user)
 /* or write fault in cacheline loop */
 105:	li	r9,1
 92:	li	r3,LG_CACHELINE_BYTES
-	b	99f
+	mfctr	r8
+	add	r0,r0,r8
+	b	106f
 /* read fault in final word loop */
 108:	li	r9,0
 	b	93f
@@ -585,7 +588,7 @@ _GLOBAL(__copy_tofrom_user)
  * r5 + (ctr << r3), and r9 is 0 for read or 1 for write.
  */
 99:	mfctr	r0
-	slw	r3,r0,r3
+106:	slw	r3,r0,r3
 	add.	r3,r3,r5
 	beq	120f			/* shouldn't happen */
 	cmpwi	0,r9,0
diff --git a/arch/ppc/mm/init.c b/arch/ppc/mm/init.c
index be02a7fec2b..363c157e361 100644
--- a/arch/ppc/mm/init.c
+++ b/arch/ppc/mm/init.c
@@ -179,6 +179,7 @@ void free_initmem(void)
 	if (!have_of)
 		FREESEC(openfirmware);
  	printk("\n");
+	ppc_md.progress = NULL;
 #undef FREESEC
 }
 
diff --git a/arch/ppc/syslib/ipic.c b/arch/ppc/syslib/ipic.c
index acb2cde3171..580ed658e87 100644
--- a/arch/ppc/syslib/ipic.c
+++ b/arch/ppc/syslib/ipic.c
@@ -479,7 +479,7 @@ void __init ipic_init(phys_addr_t phys_addr,
 	temp = 0;
 	for (i = 0 ; i < senses_count ; i++) {
 		if ((senses[i] & IRQ_SENSE_MASK) == IRQ_SENSE_EDGE) {
-			temp |= 1 << (16 - i);
+			temp |= 1 << (15 - i);
 			if (i != 0)
 				irq_desc[i + irq_offset + MPC83xx_IRQ_EXT1 - 1].status = 0;
 			else
diff --git a/arch/ppc/syslib/mpc83xx_devices.c b/arch/ppc/syslib/mpc83xx_devices.c
index 5c1a919eaab..75c8e9834ae 100644
--- a/arch/ppc/syslib/mpc83xx_devices.c
+++ b/arch/ppc/syslib/mpc83xx_devices.c
@@ -61,6 +61,7 @@ static struct plat_serial8250_port serial_platform_data[] = {
 		.iotype		= UPIO_MEM,
 		.flags		= UPF_BOOT_AUTOCONF | UPF_SKIP_TEST,
 	},
+	{ },
 };
 
 struct platform_device ppc_sys_platform_devices[] = {
diff --git a/arch/ppc/syslib/mpc85xx_devices.c b/arch/ppc/syslib/mpc85xx_devices.c
index a231795ee26..1e658ef57e7 100644
--- a/arch/ppc/syslib/mpc85xx_devices.c
+++ b/arch/ppc/syslib/mpc85xx_devices.c
@@ -61,6 +61,7 @@ static struct plat_serial8250_port serial_platform_data[] = {
 		.iotype		= UPIO_MEM,
 		.flags		= UPF_BOOT_AUTOCONF | UPF_SKIP_TEST | UPF_SHARE_IRQ,
 	},
+	{ },
 };
 
 struct platform_device ppc_sys_platform_devices[] = {
diff --git a/arch/ppc/syslib/open_pic.c b/arch/ppc/syslib/open_pic.c
index 7619e16fcca..9d4ed68b580 100644
--- a/arch/ppc/syslib/open_pic.c
+++ b/arch/ppc/syslib/open_pic.c
@@ -557,12 +557,10 @@ static void __init openpic_initipi(u_int ipi, u_int pri, u_int vec)
  */
 void openpic_cause_IPI(u_int ipi, cpumask_t cpumask)
 {
-	cpumask_t phys;
 	DECL_THIS_CPU;
 
 	CHECK_THIS_CPU;
 	check_arg_ipi(ipi);
-	phys = physmask(cpumask);
 	openpic_write(&OpenPIC->THIS_CPU.IPI_Dispatch(ipi),
 		      cpus_addr(physmask(cpumask))[0]);
 }
diff --git a/arch/ppc64/kernel/mf.c b/arch/ppc64/kernel/mf.c
index 1bd52ece497..5aca7e8005a 100644
--- a/arch/ppc64/kernel/mf.c
+++ b/arch/ppc64/kernel/mf.c
@@ -1,7 +1,7 @@
 /*
   * mf.c
   * Copyright (C) 2001 Troy D. Armstrong  IBM Corporation
-  * Copyright (C) 2004 Stephen Rothwell  IBM Corporation
+  * Copyright (C) 2004-2005 Stephen Rothwell  IBM Corporation
   *
   * This modules exists as an interface between a Linux secondary partition
   * running on an iSeries and the primary partition's Virtual Service
@@ -36,10 +36,12 @@
 
 #include <asm/time.h>
 #include <asm/uaccess.h>
+#include <asm/paca.h>
 #include <asm/iSeries/vio.h>
 #include <asm/iSeries/mf.h>
 #include <asm/iSeries/HvLpConfig.h>
 #include <asm/iSeries/ItSpCommArea.h>
+#include <asm/iSeries/ItLpQueue.h>
 
 /*
  * This is the structure layout for the Machine Facilites LPAR event
@@ -696,36 +698,23 @@ static void get_rtc_time_complete(void *token, struct ce_msg_data *ce_msg)
 	complete(&rtc->com);
 }
 
-int mf_get_rtc(struct rtc_time *tm)
+static int rtc_set_tm(int rc, u8 *ce_msg, struct rtc_time *tm)
 {
-	struct ce_msg_comp_data ce_complete;
-	struct rtc_time_data rtc_data;
-	int rc;
-
-	memset(&ce_complete, 0, sizeof(ce_complete));
-	memset(&rtc_data, 0, sizeof(rtc_data));
-	init_completion(&rtc_data.com);
-	ce_complete.handler = &get_rtc_time_complete;
-	ce_complete.token = &rtc_data;
-	rc = signal_ce_msg_simple(0x40, &ce_complete);
-	if (rc)
-		return rc;
-	wait_for_completion(&rtc_data.com);
 	tm->tm_wday = 0;
 	tm->tm_yday = 0;
 	tm->tm_isdst = 0;
-	if (rtc_data.rc) {
+	if (rc) {
 		tm->tm_sec = 0;
 		tm->tm_min = 0;
 		tm->tm_hour = 0;
 		tm->tm_mday = 15;
 		tm->tm_mon = 5;
 		tm->tm_year = 52;
-		return rtc_data.rc;
+		return rc;
 	}
 
-	if ((rtc_data.ce_msg.ce_msg[2] == 0xa9) ||
-	    (rtc_data.ce_msg.ce_msg[2] == 0xaf)) {
+	if ((ce_msg[2] == 0xa9) ||
+	    (ce_msg[2] == 0xaf)) {
 		/* TOD clock is not set */
 		tm->tm_sec = 1;
 		tm->tm_min = 1;
@@ -736,7 +725,6 @@ int mf_get_rtc(struct rtc_time *tm)
 		mf_set_rtc(tm);
 	}
 	{
-		u8 *ce_msg = rtc_data.ce_msg.ce_msg;
 		u8 year = ce_msg[5];
 		u8 sec = ce_msg[6];
 		u8 min = ce_msg[7];
@@ -765,6 +753,63 @@ int mf_get_rtc(struct rtc_time *tm)
 	return 0;
 }
 
+int mf_get_rtc(struct rtc_time *tm)
+{
+	struct ce_msg_comp_data ce_complete;
+	struct rtc_time_data rtc_data;
+	int rc;
+
+	memset(&ce_complete, 0, sizeof(ce_complete));
+	memset(&rtc_data, 0, sizeof(rtc_data));
+	init_completion(&rtc_data.com);
+	ce_complete.handler = &get_rtc_time_complete;
+	ce_complete.token = &rtc_data;
+	rc = signal_ce_msg_simple(0x40, &ce_complete);
+	if (rc)
+		return rc;
+	wait_for_completion(&rtc_data.com);
+	return rtc_set_tm(rtc_data.rc, rtc_data.ce_msg.ce_msg, tm);
+}
+
+struct boot_rtc_time_data {
+	int busy;
+	struct ce_msg_data ce_msg;
+	int rc;
+};
+
+static void get_boot_rtc_time_complete(void *token, struct ce_msg_data *ce_msg)
+{
+	struct boot_rtc_time_data *rtc = token;
+
+	memcpy(&rtc->ce_msg, ce_msg, sizeof(rtc->ce_msg));
+	rtc->rc = 0;
+	rtc->busy = 0;
+}
+
+int mf_get_boot_rtc(struct rtc_time *tm)
+{
+	struct ce_msg_comp_data ce_complete;
+	struct boot_rtc_time_data rtc_data;
+	int rc;
+
+	memset(&ce_complete, 0, sizeof(ce_complete));
+	memset(&rtc_data, 0, sizeof(rtc_data));
+	rtc_data.busy = 1;
+	ce_complete.handler = &get_boot_rtc_time_complete;
+	ce_complete.token = &rtc_data;
+	rc = signal_ce_msg_simple(0x40, &ce_complete);
+	if (rc)
+		return rc;
+	/* We need to poll here as we are not yet taking interrupts */
+	while (rtc_data.busy) {
+		extern unsigned long lpevent_count;
+		struct ItLpQueue *lpq = get_paca()->lpqueue_ptr;
+		if (lpq && ItLpQueue_isLpIntPending(lpq))
+			lpevent_count += ItLpQueue_process(lpq, NULL);
+	}
+	return rtc_set_tm(rtc_data.rc, rtc_data.ce_msg.ce_msg, tm);
+}
+
 int mf_set_rtc(struct rtc_time *tm)
 {
 	char ce_time[12];
diff --git a/arch/ppc64/kernel/pmac_smp.c b/arch/ppc64/kernel/pmac_smp.c
index c27588ede2f..a23de37227b 100644
--- a/arch/ppc64/kernel/pmac_smp.c
+++ b/arch/ppc64/kernel/pmac_smp.c
@@ -68,6 +68,7 @@ extern struct smp_ops_t *smp_ops;
 
 static void (*pmac_tb_freeze)(int freeze);
 static struct device_node *pmac_tb_clock_chip_host;
+static u8 pmac_tb_pulsar_addr;
 static DEFINE_SPINLOCK(timebase_lock);
 static unsigned long timebase;
 
@@ -106,12 +107,9 @@ static void smp_core99_pulsar_tb_freeze(int freeze)
 	u8 data;
 	int rc;
 
-	/* Strangely, the device-tree says address is 0xd2, but darwin
-	 * accesses 0xd0 ...
-	 */
 	pmac_low_i2c_setmode(pmac_tb_clock_chip_host, pmac_low_i2c_mode_combined);
 	rc = pmac_low_i2c_xfer(pmac_tb_clock_chip_host,
-			       0xd4 | pmac_low_i2c_read,
+			       pmac_tb_pulsar_addr | pmac_low_i2c_read,
 			       0x2e, &data, 1);
 	if (rc != 0)
 		goto bail;
@@ -120,7 +118,7 @@ static void smp_core99_pulsar_tb_freeze(int freeze)
 
 	pmac_low_i2c_setmode(pmac_tb_clock_chip_host, pmac_low_i2c_mode_stdsub);
 	rc = pmac_low_i2c_xfer(pmac_tb_clock_chip_host,
-			       0xd4 | pmac_low_i2c_write,
+			       pmac_tb_pulsar_addr | pmac_low_i2c_write,
 			       0x2e, &data, 1);
  bail:
 	if (rc != 0) {
@@ -185,6 +183,12 @@ static int __init smp_core99_probe(void)
 	if (ncpus <= 1)
 		return 1;
 
+	/* HW sync only on these platforms */
+	if (!machine_is_compatible("PowerMac7,2") &&
+	    !machine_is_compatible("PowerMac7,3") &&
+	    !machine_is_compatible("RackMac3,1"))
+		goto nohwsync;
+
 	/* Look for the clock chip */
 	for (cc = NULL; (cc = of_find_node_by_name(cc, "i2c-hwclock")) != NULL;) {
 		struct device_node *p = of_get_parent(cc);
@@ -198,11 +202,18 @@ static int __init smp_core99_probe(void)
 			goto next;
 		switch (*reg) {
 		case 0xd2:
-			pmac_tb_freeze = smp_core99_cypress_tb_freeze;
-			printk(KERN_INFO "Timebase clock is Cypress chip\n");
+			if (device_is_compatible(cc, "pulsar-legacy-slewing")) {
+				pmac_tb_freeze = smp_core99_pulsar_tb_freeze;
+				pmac_tb_pulsar_addr = 0xd2;
+				printk(KERN_INFO "Timebase clock is Pulsar chip\n");
+			} else if (device_is_compatible(cc, "cy28508")) {
+				pmac_tb_freeze = smp_core99_cypress_tb_freeze;
+				printk(KERN_INFO "Timebase clock is Cypress chip\n");
+			}
 			break;
 		case 0xd4:
 			pmac_tb_freeze = smp_core99_pulsar_tb_freeze;
+			pmac_tb_pulsar_addr = 0xd4;
 			printk(KERN_INFO "Timebase clock is Pulsar chip\n");
 			break;
 		}
@@ -210,12 +221,15 @@ static int __init smp_core99_probe(void)
 			pmac_tb_clock_chip_host = p;
 			smp_ops->give_timebase = smp_core99_give_timebase;
 			smp_ops->take_timebase = smp_core99_take_timebase;
+			of_node_put(cc);
+			of_node_put(p);
 			break;
 		}
 	next:
 		of_node_put(p);
 	}
 
+ nohwsync:
 	mpic_request_ipis();
 
 	return ncpus;
diff --git a/arch/ppc64/kernel/prom_init.c b/arch/ppc64/kernel/prom_init.c
index 35ec42de962..6f79b7b9b44 100644
--- a/arch/ppc64/kernel/prom_init.c
+++ b/arch/ppc64/kernel/prom_init.c
@@ -1750,7 +1750,44 @@ static void __init flatten_device_tree(void)
 	prom_printf("Device tree struct  0x%x -> 0x%x\n",
 		    RELOC(dt_struct_start), RELOC(dt_struct_end));
 
- }
+}
+
+
+static void __init fixup_device_tree(void)
+{
+	unsigned long offset = reloc_offset();
+	phandle u3, i2c, mpic;
+	u32 u3_rev;
+	u32 interrupts[2];
+	u32 parent;
+
+	/* Some G5s have a missing interrupt definition, fix it up here */
+	u3 = call_prom("finddevice", 1, 1, ADDR("/u3@0,f8000000"));
+	if ((long)u3 <= 0)
+		return;
+	i2c = call_prom("finddevice", 1, 1, ADDR("/u3@0,f8000000/i2c@f8001000"));
+	if ((long)i2c <= 0)
+		return;
+	mpic = call_prom("finddevice", 1, 1, ADDR("/u3@0,f8000000/mpic@f8040000"));
+	if ((long)mpic <= 0)
+		return;
+
+	/* check if proper rev of u3 */
+	if (prom_getprop(u3, "device-rev", &u3_rev, sizeof(u3_rev)) <= 0)
+		return;
+	if (u3_rev != 0x35)
+		return;
+	/* does it need fixup ? */
+	if (prom_getproplen(i2c, "interrupts") > 0)
+		return;
+	/* interrupt on this revision of u3 is number 0 and level */
+	interrupts[0] = 0;
+	interrupts[1] = 1;
+	prom_setprop(i2c, "interrupts", &interrupts, sizeof(interrupts));
+	parent = (u32)mpic;
+	prom_setprop(i2c, "interrupt-parent", &parent, sizeof(parent));
+}
+
 
 static void __init prom_find_boot_cpu(void)
 {
@@ -1920,6 +1957,11 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4, unsigned long
 	}
 
 	/*
+	 * Fixup any known bugs in the device-tree
+	 */
+	fixup_device_tree();
+
+	/*
 	 * Now finally create the flattened device-tree
 	 */
        	prom_printf("copying OF device tree ...\n");
diff --git a/arch/ppc64/kernel/rtc.c b/arch/ppc64/kernel/rtc.c
index 3e70b91375f..67989055a9f 100644
--- a/arch/ppc64/kernel/rtc.c
+++ b/arch/ppc64/kernel/rtc.c
@@ -292,47 +292,10 @@ int iSeries_set_rtc_time(struct rtc_time *tm)
 
 void iSeries_get_boot_time(struct rtc_time *tm)
 {
-	unsigned long time;
-	static unsigned long lastsec = 1;
-
-	u32 dataWord1 = *((u32 *)(&xSpCommArea.xBcdTimeAtIplStart));
-	u32 dataWord2 = *(((u32 *)&(xSpCommArea.xBcdTimeAtIplStart)) + 1);
-	int year = 1970;
-	int year1 = ( dataWord1 >> 24 ) & 0x000000FF;
-	int year2 = ( dataWord1 >> 16 ) & 0x000000FF;
-	int sec = ( dataWord1 >> 8 ) & 0x000000FF;
-	int min = dataWord1 & 0x000000FF;
-	int hour = ( dataWord2 >> 24 ) & 0x000000FF;
-	int day = ( dataWord2 >> 8 ) & 0x000000FF;
-	int mon = dataWord2 & 0x000000FF;
-
 	if ( piranha_simulator )
 		return;
 
-	BCD_TO_BIN(sec);
-	BCD_TO_BIN(min);
-	BCD_TO_BIN(hour);
-	BCD_TO_BIN(day);
-	BCD_TO_BIN(mon);
-	BCD_TO_BIN(year1);
-	BCD_TO_BIN(year2);
-	year = year1 * 100 + year2;
-
-	time = mktime(year, mon, day, hour, min, sec);
-	time += ( jiffies / HZ );
-
-	/* Now THIS is a nasty hack!
-	* It ensures that the first two calls get different answers.  
-	* That way the loop in init_time (time.c) will not think
-	* the clock is stuck.
-	*/
-	if ( lastsec ) {
-		time -= lastsec;
-		--lastsec;
-	}
-
-	to_tm(time, tm); 
-	tm->tm_year -= 1900;
+	mf_get_boot_rtc(tm);
 	tm->tm_mon  -= 1;
 }
 #endif
diff --git a/arch/ppc64/kernel/time.c b/arch/ppc64/kernel/time.c
index 772a465b49f..3d54745108c 100644
--- a/arch/ppc64/kernel/time.c
+++ b/arch/ppc64/kernel/time.c
@@ -515,6 +515,7 @@ void __init time_init(void)
 	do_gtod.varp = &do_gtod.vars[0];
 	do_gtod.var_idx = 0;
 	do_gtod.varp->tb_orig_stamp = tb_last_stamp;
+	get_paca()->next_jiffy_update_tb = tb_last_stamp + tb_ticks_per_jiffy;
 	do_gtod.varp->stamp_xsec = xtime.tv_sec * XSEC_PER_SEC;
 	do_gtod.tb_ticks_per_sec = tb_ticks_per_sec;
 	do_gtod.varp->tb_to_xs = tb_to_xs;
diff --git a/arch/sparc64/kernel/pci_iommu.c b/arch/sparc64/kernel/pci_iommu.c
index 292983413ae..33ca56c90da 100644
--- a/arch/sparc64/kernel/pci_iommu.c
+++ b/arch/sparc64/kernel/pci_iommu.c
@@ -8,6 +8,7 @@
 #include <linux/kernel.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
+#include <linux/delay.h>
 
 #include <asm/pbm.h>
 
@@ -379,6 +380,56 @@ bad:
 	return PCI_DMA_ERROR_CODE;
 }
 
+static void pci_strbuf_flush(struct pci_strbuf *strbuf, struct pci_iommu *iommu, u32 vaddr, unsigned long ctx, unsigned long npages)
+{
+	int limit;
+
+	PCI_STC_FLUSHFLAG_INIT(strbuf);
+	if (strbuf->strbuf_ctxflush &&
+	    iommu->iommu_ctxflush) {
+		unsigned long matchreg, flushreg;
+
+		flushreg = strbuf->strbuf_ctxflush;
+		matchreg = PCI_STC_CTXMATCH_ADDR(strbuf, ctx);
+
+		limit = 100000;
+		pci_iommu_write(flushreg, ctx);
+		for(;;) {
+			if (((long)pci_iommu_read(matchreg)) >= 0L)
+				break;
+			limit--;
+			if (!limit)
+				break;
+			udelay(1);
+		}
+		if (!limit)
+			printk(KERN_WARNING "pci_strbuf_flush: ctx flush "
+			       "timeout vaddr[%08x] ctx[%lx]\n",
+			       vaddr, ctx);
+	} else {
+		unsigned long i;
+
+		for (i = 0; i < npages; i++, vaddr += IO_PAGE_SIZE)
+			pci_iommu_write(strbuf->strbuf_pflush, vaddr);
+	}
+
+	pci_iommu_write(strbuf->strbuf_fsync, strbuf->strbuf_flushflag_pa);
+	(void) pci_iommu_read(iommu->write_complete_reg);
+
+	limit = 100000;
+	while (!PCI_STC_FLUSHFLAG_SET(strbuf)) {
+		limit--;
+		if (!limit)
+			break;
+		udelay(1);
+		membar("#LoadLoad");
+	}
+	if (!limit)
+		printk(KERN_WARNING "pci_strbuf_flush: flushflag timeout "
+		       "vaddr[%08x] ctx[%lx] npages[%ld]\n",
+		       vaddr, ctx, npages);
+}
+
 /* Unmap a single streaming mode DMA translation. */
 void pci_unmap_single(struct pci_dev *pdev, dma_addr_t bus_addr, size_t sz, int direction)
 {
@@ -386,7 +437,7 @@ void pci_unmap_single(struct pci_dev *pdev, dma_addr_t bus_addr, size_t sz, int
 	struct pci_iommu *iommu;
 	struct pci_strbuf *strbuf;
 	iopte_t *base;
-	unsigned long flags, npages, i, ctx;
+	unsigned long flags, npages, ctx;
 
 	if (direction == PCI_DMA_NONE)
 		BUG();
@@ -414,29 +465,8 @@ void pci_unmap_single(struct pci_dev *pdev, dma_addr_t bus_addr, size_t sz, int
 		ctx = (iopte_val(*base) & IOPTE_CONTEXT) >> 47UL;
 
 	/* Step 1: Kick data out of streaming buffers if necessary. */
-	if (strbuf->strbuf_enabled) {
-		u32 vaddr = bus_addr;
-
-		PCI_STC_FLUSHFLAG_INIT(strbuf);
-		if (strbuf->strbuf_ctxflush &&
-		    iommu->iommu_ctxflush) {
-			unsigned long matchreg, flushreg;
-
-			flushreg = strbuf->strbuf_ctxflush;
-			matchreg = PCI_STC_CTXMATCH_ADDR(strbuf, ctx);
-			do {
-				pci_iommu_write(flushreg, ctx);
-			} while(((long)pci_iommu_read(matchreg)) < 0L);
-		} else {
-			for (i = 0; i < npages; i++, vaddr += IO_PAGE_SIZE)
-				pci_iommu_write(strbuf->strbuf_pflush, vaddr);
-		}
-
-		pci_iommu_write(strbuf->strbuf_fsync, strbuf->strbuf_flushflag_pa);
-		(void) pci_iommu_read(iommu->write_complete_reg);
-		while (!PCI_STC_FLUSHFLAG_SET(strbuf))
-			membar("#LoadLoad");
-	}
+	if (strbuf->strbuf_enabled)
+		pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, npages);
 
 	/* Step 2: Clear out first TSB entry. */
 	iopte_make_dummy(iommu, base);
@@ -647,29 +677,8 @@ void pci_unmap_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems,
 		ctx = (iopte_val(*base) & IOPTE_CONTEXT) >> 47UL;
 
 	/* Step 1: Kick data out of streaming buffers if necessary. */
-	if (strbuf->strbuf_enabled) {
-		u32 vaddr = (u32) bus_addr;
-
-		PCI_STC_FLUSHFLAG_INIT(strbuf);
-		if (strbuf->strbuf_ctxflush &&
-		    iommu->iommu_ctxflush) {
-			unsigned long matchreg, flushreg;
-
-			flushreg = strbuf->strbuf_ctxflush;
-			matchreg = PCI_STC_CTXMATCH_ADDR(strbuf, ctx);
-			do {
-				pci_iommu_write(flushreg, ctx);
-			} while(((long)pci_iommu_read(matchreg)) < 0L);
-		} else {
-			for (i = 0; i < npages; i++, vaddr += IO_PAGE_SIZE)
-				pci_iommu_write(strbuf->strbuf_pflush, vaddr);
-		}
-
-		pci_iommu_write(strbuf->strbuf_fsync, strbuf->strbuf_flushflag_pa);
-		(void) pci_iommu_read(iommu->write_complete_reg);
-		while (!PCI_STC_FLUSHFLAG_SET(strbuf))
-			membar("#LoadLoad");
-	}
+	if (strbuf->strbuf_enabled)
+		pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, npages);
 
 	/* Step 2: Clear out first TSB entry. */
 	iopte_make_dummy(iommu, base);
@@ -715,28 +724,7 @@ void pci_dma_sync_single_for_cpu(struct pci_dev *pdev, dma_addr_t bus_addr, size
 	}
 
 	/* Step 2: Kick data out of streaming buffers. */
-	PCI_STC_FLUSHFLAG_INIT(strbuf);
-	if (iommu->iommu_ctxflush &&
-	    strbuf->strbuf_ctxflush) {
-		unsigned long matchreg, flushreg;
-
-		flushreg = strbuf->strbuf_ctxflush;
-		matchreg = PCI_STC_CTXMATCH_ADDR(strbuf, ctx);
-		do {
-			pci_iommu_write(flushreg, ctx);
-		} while(((long)pci_iommu_read(matchreg)) < 0L);
-	} else {
-		unsigned long i;
-
-		for (i = 0; i < npages; i++, bus_addr += IO_PAGE_SIZE)
-			pci_iommu_write(strbuf->strbuf_pflush, bus_addr);
-	}
-
-	/* Step 3: Perform flush synchronization sequence. */
-	pci_iommu_write(strbuf->strbuf_fsync, strbuf->strbuf_flushflag_pa);
-	(void) pci_iommu_read(iommu->write_complete_reg);
-	while (!PCI_STC_FLUSHFLAG_SET(strbuf))
-		membar("#LoadLoad");
+	pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, npages);
 
 	spin_unlock_irqrestore(&iommu->lock, flags);
 }
@@ -749,7 +737,8 @@ void pci_dma_sync_sg_for_cpu(struct pci_dev *pdev, struct scatterlist *sglist, i
 	struct pcidev_cookie *pcp;
 	struct pci_iommu *iommu;
 	struct pci_strbuf *strbuf;
-	unsigned long flags, ctx;
+	unsigned long flags, ctx, npages, i;
+	u32 bus_addr;
 
 	pcp = pdev->sysdata;
 	iommu = pcp->pbm->iommu;
@@ -772,36 +761,14 @@ void pci_dma_sync_sg_for_cpu(struct pci_dev *pdev, struct scatterlist *sglist, i
 	}
 
 	/* Step 2: Kick data out of streaming buffers. */
-	PCI_STC_FLUSHFLAG_INIT(strbuf);
-	if (iommu->iommu_ctxflush &&
-	    strbuf->strbuf_ctxflush) {
-		unsigned long matchreg, flushreg;
-
-		flushreg = strbuf->strbuf_ctxflush;
-		matchreg = PCI_STC_CTXMATCH_ADDR(strbuf, ctx);
-		do {
-			pci_iommu_write(flushreg, ctx);
-		} while (((long)pci_iommu_read(matchreg)) < 0L);
-	} else {
-		unsigned long i, npages;
-		u32 bus_addr;
-
-		bus_addr = sglist[0].dma_address & IO_PAGE_MASK;
-
-		for(i = 1; i < nelems; i++)
-			if (!sglist[i].dma_length)
-				break;
-		i--;
-		npages = (IO_PAGE_ALIGN(sglist[i].dma_address + sglist[i].dma_length) - bus_addr) >> IO_PAGE_SHIFT;
-		for (i = 0; i < npages; i++, bus_addr += IO_PAGE_SIZE)
-			pci_iommu_write(strbuf->strbuf_pflush, bus_addr);
-	}
-
-	/* Step 3: Perform flush synchronization sequence. */
-	pci_iommu_write(strbuf->strbuf_fsync, strbuf->strbuf_flushflag_pa);
-	(void) pci_iommu_read(iommu->write_complete_reg);
-	while (!PCI_STC_FLUSHFLAG_SET(strbuf))
-		membar("#LoadLoad");
+	bus_addr = sglist[0].dma_address & IO_PAGE_MASK;
+	for(i = 1; i < nelems; i++)
+		if (!sglist[i].dma_length)
+			break;
+	i--;
+	npages = (IO_PAGE_ALIGN(sglist[i].dma_address + sglist[i].dma_length)
+		  - bus_addr) >> IO_PAGE_SHIFT;
+	pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, npages);
 
 	spin_unlock_irqrestore(&iommu->lock, flags);
 }
diff --git a/arch/sparc64/kernel/sbus.c b/arch/sparc64/kernel/sbus.c
index 14d9c3a21b9..76ea6455433 100644
--- a/arch/sparc64/kernel/sbus.c
+++ b/arch/sparc64/kernel/sbus.c
@@ -117,19 +117,34 @@ static void iommu_flush(struct sbus_iommu *iommu, u32 base, unsigned long npages
 
 #define STRBUF_TAG_VALID	0x02UL
 
-static void strbuf_flush(struct sbus_iommu *iommu, u32 base, unsigned long npages)
+static void sbus_strbuf_flush(struct sbus_iommu *iommu, u32 base, unsigned long npages)
 {
+	unsigned long n;
+	int limit;
+
 	iommu->strbuf_flushflag = 0UL;
-	while (npages--)
-		upa_writeq(base + (npages << IO_PAGE_SHIFT),
+	n = npages;
+	while (n--)
+		upa_writeq(base + (n << IO_PAGE_SHIFT),
 			   iommu->strbuf_regs + STRBUF_PFLUSH);
 
 	/* Whoopee cushion! */
 	upa_writeq(__pa(&iommu->strbuf_flushflag),
 		   iommu->strbuf_regs + STRBUF_FSYNC);
 	upa_readq(iommu->sbus_control_reg);
-	while (iommu->strbuf_flushflag == 0UL)
+
+	limit = 100000;
+	while (iommu->strbuf_flushflag == 0UL) {
+		limit--;
+		if (!limit)
+			break;
+		udelay(1);
 		membar("#LoadLoad");
+	}
+	if (!limit)
+		printk(KERN_WARNING "sbus_strbuf_flush: flushflag timeout "
+		       "vaddr[%08x] npages[%ld]\n",
+		       base, npages);
 }
 
 static iopte_t *alloc_streaming_cluster(struct sbus_iommu *iommu, unsigned long npages)
@@ -406,7 +421,7 @@ void sbus_unmap_single(struct sbus_dev *sdev, dma_addr_t dma_addr, size_t size,
 
 	spin_lock_irqsave(&iommu->lock, flags);
 	free_streaming_cluster(iommu, dma_base, size >> IO_PAGE_SHIFT);
-	strbuf_flush(iommu, dma_base, size >> IO_PAGE_SHIFT);
+	sbus_strbuf_flush(iommu, dma_base, size >> IO_PAGE_SHIFT);
 	spin_unlock_irqrestore(&iommu->lock, flags);
 }
 
@@ -569,7 +584,7 @@ void sbus_unmap_sg(struct sbus_dev *sdev, struct scatterlist *sg, int nents, int
 	iommu = sdev->bus->iommu;
 	spin_lock_irqsave(&iommu->lock, flags);
 	free_streaming_cluster(iommu, dvma_base, size >> IO_PAGE_SHIFT);
-	strbuf_flush(iommu, dvma_base, size >> IO_PAGE_SHIFT);
+	sbus_strbuf_flush(iommu, dvma_base, size >> IO_PAGE_SHIFT);
 	spin_unlock_irqrestore(&iommu->lock, flags);
 }
 
@@ -581,7 +596,7 @@ void sbus_dma_sync_single_for_cpu(struct sbus_dev *sdev, dma_addr_t base, size_t
 	size = (IO_PAGE_ALIGN(base + size) - (base & IO_PAGE_MASK));
 
 	spin_lock_irqsave(&iommu->lock, flags);
-	strbuf_flush(iommu, base & IO_PAGE_MASK, size >> IO_PAGE_SHIFT);
+	sbus_strbuf_flush(iommu, base & IO_PAGE_MASK, size >> IO_PAGE_SHIFT);
 	spin_unlock_irqrestore(&iommu->lock, flags);
 }
 
@@ -605,7 +620,7 @@ void sbus_dma_sync_sg_for_cpu(struct sbus_dev *sdev, struct scatterlist *sg, int
 	size = IO_PAGE_ALIGN(sg[i].dma_address + sg[i].dma_length) - base;
 
 	spin_lock_irqsave(&iommu->lock, flags);
-	strbuf_flush(iommu, base, size >> IO_PAGE_SHIFT);
+	sbus_strbuf_flush(iommu, base, size >> IO_PAGE_SHIFT);
 	spin_unlock_irqrestore(&iommu->lock, flags);
 }
 
diff --git a/arch/sparc64/kernel/setup.c b/arch/sparc64/kernel/setup.c
index 12c3d84b746..b7e6a91952b 100644
--- a/arch/sparc64/kernel/setup.c
+++ b/arch/sparc64/kernel/setup.c
@@ -383,6 +383,17 @@ static void __init process_switch(char c)
 		/* Use PROM debug console. */
 		register_console(&prom_debug_console);
 		break;
+	case 'P':
+		/* Force UltraSPARC-III P-Cache on. */
+		if (tlb_type != cheetah) {
+			printk("BOOT: Ignoring P-Cache force option.\n");
+			break;
+		}
+		cheetah_pcache_forced_on = 1;
+		add_taint(TAINT_MACHINE_CHECK);
+		cheetah_enable_pcache();
+		break;
+
 	default:
 		printk("Unknown boot switch (-%c)\n", c);
 		break;
diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c
index 6dff06a44e7..e5b9c7a2778 100644
--- a/arch/sparc64/kernel/smp.c
+++ b/arch/sparc64/kernel/smp.c
@@ -123,6 +123,9 @@ void __init smp_callin(void)
 
 	smp_setup_percpu_timer();
 
+	if (cheetah_pcache_forced_on)
+		cheetah_enable_pcache();
+
 	local_irq_enable();
 
 	calibrate_delay();
diff --git a/arch/sparc64/kernel/traps.c b/arch/sparc64/kernel/traps.c
index 56b203a2af6..a9f4596d7c2 100644
--- a/arch/sparc64/kernel/traps.c
+++ b/arch/sparc64/kernel/traps.c
@@ -421,6 +421,25 @@ asmlinkage void cee_log(unsigned long ce_status,
 	}
 }
 
+int cheetah_pcache_forced_on;
+
+void cheetah_enable_pcache(void)
+{
+	unsigned long dcr;
+
+	printk("CHEETAH: Enabling P-Cache on cpu %d.\n",
+	       smp_processor_id());
+
+	__asm__ __volatile__("ldxa [%%g0] %1, %0"
+			     : "=r" (dcr)
+			     : "i" (ASI_DCU_CONTROL_REG));
+	dcr |= (DCU_PE | DCU_HPE | DCU_SPE | DCU_SL);
+	__asm__ __volatile__("stxa %0, [%%g0] %1\n\t"
+			     "membar #Sync"
+			     : /* no outputs */
+			     : "r" (dcr), "i" (ASI_DCU_CONTROL_REG));
+}
+
 /* Cheetah error trap handling. */
 static unsigned long ecache_flush_physbase;
 static unsigned long ecache_flush_linesize;
diff --git a/arch/um/Kconfig_x86_64 b/arch/um/Kconfig_x86_64
index fd8d7e8982b..f162f50f0b1 100644
--- a/arch/um/Kconfig_x86_64
+++ b/arch/um/Kconfig_x86_64
@@ -6,6 +6,10 @@ config 64BIT
 	bool
 	default y
 
+config TOP_ADDR
+ 	hex
+	default 0x80000000
+
 config 3_LEVEL_PGTABLES
        bool
        default y
diff --git a/arch/um/drivers/chan_kern.c b/arch/um/drivers/chan_kern.c
index 0150038af79..14a12d6b3df 100644
--- a/arch/um/drivers/chan_kern.c
+++ b/arch/um/drivers/chan_kern.c
@@ -20,9 +20,17 @@
 #include "os.h"
 
 #ifdef CONFIG_NOCONFIG_CHAN
+
+/* The printk's here are wrong because we are complaining that there is no
+ * output device, but printk is printing to that output device.  The user will
+ * never see the error.  printf would be better, except it can't run on a
+ * kernel stack because it will overflow it.
+ * Use printk for now since that will avoid crashing.
+ */
+
 static void *not_configged_init(char *str, int device, struct chan_opts *opts)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 	return(NULL);
 }
@@ -30,27 +38,27 @@ static void *not_configged_init(char *str, int device, struct chan_opts *opts)
 static int not_configged_open(int input, int output, int primary, void *data,
 			      char **dev_out)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 	return(-ENODEV);
 }
 
 static void not_configged_close(int fd, void *data)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 }
 
 static int not_configged_read(int fd, char *c_out, void *data)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 	return(-EIO);
 }
 
 static int not_configged_write(int fd, const char *buf, int len, void *data)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 	return(-EIO);
 }
@@ -58,7 +66,7 @@ static int not_configged_write(int fd, const char *buf, int len, void *data)
 static int not_configged_console_write(int fd, const char *buf, int len,
 				       void *data)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 	return(-EIO);
 }
@@ -66,7 +74,7 @@ static int not_configged_console_write(int fd, const char *buf, int len,
 static int not_configged_window_size(int fd, void *data, unsigned short *rows,
 				     unsigned short *cols)
 {
-	printf(KERN_ERR "Using a channel type which is configured out of "
+	printk(KERN_ERR "Using a channel type which is configured out of "
 	       "UML\n");
 	return(-ENODEV);
 }
diff --git a/arch/um/drivers/mcast_kern.c b/arch/um/drivers/mcast_kern.c
index faf714e87b5..217438cdef3 100644
--- a/arch/um/drivers/mcast_kern.c
+++ b/arch/um/drivers/mcast_kern.c
@@ -73,7 +73,6 @@ int mcast_setup(char *str, char **mac_out, void *data)
 	struct mcast_init *init = data;
 	char *port_str = NULL, *ttl_str = NULL, *remain;
 	char *last;
-	int n;
 
 	*init = ((struct mcast_init)
 		{ .addr 	= "239.192.168.1",
@@ -89,13 +88,12 @@ int mcast_setup(char *str, char **mac_out, void *data)
 	}
 	
 	if(port_str != NULL){
-		n = simple_strtoul(port_str, &last, 10);
+		init->port = simple_strtoul(port_str, &last, 10);
 		if((*last != '\0') || (last == port_str)){
 			printk(KERN_ERR "mcast_setup - Bad port : '%s'\n", 
 			       port_str);
 			return(0);
 		}
-		init->port = htons(n);
 	}
 
 	if(ttl_str != NULL){
diff --git a/arch/um/drivers/mcast_user.c b/arch/um/drivers/mcast_user.c
index 0fe1d9fa913..7a0d115b29d 100644
--- a/arch/um/drivers/mcast_user.c
+++ b/arch/um/drivers/mcast_user.c
@@ -38,7 +38,7 @@ static struct sockaddr_in *new_addr(char *addr, unsigned short port)
 	}
 	sin->sin_family = AF_INET;
 	sin->sin_addr.s_addr = in_aton(addr);
-	sin->sin_port = port;
+	sin->sin_port = htons(port);
 	return(sin);
 }
 
@@ -55,28 +55,25 @@ static int mcast_open(void *data)
 	struct mcast_data *pri = data;
 	struct sockaddr_in *sin = pri->mcast_addr;
 	struct ip_mreq mreq;
-	int fd, yes = 1;
+	int fd = -EINVAL, yes = 1, err = -EINVAL;;
 
 
-	if ((sin->sin_addr.s_addr == 0) || (sin->sin_port == 0)) {
-		fd = -EINVAL;
+	if ((sin->sin_addr.s_addr == 0) || (sin->sin_port == 0))
 		goto out;
-	}
 
 	fd = socket(AF_INET, SOCK_DGRAM, 0);
+
 	if (fd < 0){
 		printk("mcast_open : data socket failed, errno = %d\n", 
 		       errno);
-		fd = -ENOMEM;
+		fd = -errno;
 		goto out;
 	}
 
 	if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &yes, sizeof(yes)) < 0) {
 		printk("mcast_open: SO_REUSEADDR failed, errno = %d\n",
 			errno);
-		os_close_file(fd);
-		fd = -EINVAL;
-		goto out;
+		goto out_close;
 	}
 
 	/* set ttl according to config */
@@ -84,26 +81,20 @@ static int mcast_open(void *data)
 		       sizeof(pri->ttl)) < 0) {
 		printk("mcast_open: IP_MULTICAST_TTL failed, error = %d\n",
 			errno);
-		os_close_file(fd);
-		fd = -EINVAL;
-		goto out;
+		goto out_close;
 	}
 
 	/* set LOOP, so data does get fed back to local sockets */
 	if (setsockopt(fd, SOL_IP, IP_MULTICAST_LOOP, &yes, sizeof(yes)) < 0) {
 		printk("mcast_open: IP_MULTICAST_LOOP failed, error = %d\n",
 			errno);
-		os_close_file(fd);
-		fd = -EINVAL;
-		goto out;
+		goto out_close;
 	}
 
 	/* bind socket to mcast address */
 	if (bind(fd, (struct sockaddr *) sin, sizeof(*sin)) < 0) {
 		printk("mcast_open : data bind failed, errno = %d\n", errno);
-		os_close_file(fd);
-		fd = -EINVAL;
-		goto out;
+		goto out_close;
 	}		
 	
 	/* subscribe to the multicast group */
@@ -117,12 +108,15 @@ static int mcast_open(void *data)
 		       "interface on the host.\n");
 		printk("eth0 should be configured in order to use the "
 		       "multicast transport.\n");
-		os_close_file(fd);
-		fd = -EINVAL;
+                goto out_close;
 	}
 
  out:
-	return(fd);
+	return fd;
+
+ out_close:
+        os_close_file(fd);
+        return err;
 }
 
 static void mcast_close(int fd, void *data)
@@ -164,14 +158,3 @@ struct net_user_info mcast_user_info = {
 	.delete_address = NULL,
 	.max_packet	= MAX_PACKET - ETH_HEADER_OTHER
 };
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
diff --git a/arch/um/drivers/ubd_kern.c b/arch/um/drivers/ubd_kern.c
index 9a56ff94308..88f956c34fe 100644
--- a/arch/um/drivers/ubd_kern.c
+++ b/arch/um/drivers/ubd_kern.c
@@ -55,7 +55,7 @@
 #include "mem_kern.h"
 #include "cow.h"
 
-enum ubd_req { UBD_READ, UBD_WRITE, UBD_MMAP };
+enum ubd_req { UBD_READ, UBD_WRITE };
 
 struct io_thread_req {
 	enum ubd_req op;
@@ -68,8 +68,6 @@ struct io_thread_req {
 	unsigned long sector_mask;
 	unsigned long long cow_offset;
 	unsigned long bitmap_words[2];
-	int map_fd;
-	unsigned long long map_offset;
 	int error;
 };
 
@@ -122,10 +120,6 @@ static int ubd_ioctl(struct inode * inode, struct file * file,
 
 #define MAX_DEV (8)
 
-/* Changed in early boot */
-static int ubd_do_mmap = 0;
-#define UBD_MMAP_BLOCK_SIZE PAGE_SIZE
-
 static struct block_device_operations ubd_blops = {
         .owner		= THIS_MODULE,
         .open		= ubd_open,
@@ -175,12 +169,6 @@ struct ubd {
 	int no_cow;
 	struct cow cow;
 	struct platform_device pdev;
-
-	int map_writes;
-	int map_reads;
-	int nomap_writes;
-	int nomap_reads;
-	int write_maps;
 };
 
 #define DEFAULT_COW { \
@@ -200,11 +188,6 @@ struct ubd {
 	.openflags =		OPEN_FLAGS, \
         .no_cow =               0, \
         .cow =			DEFAULT_COW, \
-	.map_writes		= 0, \
-	.map_reads		= 0, \
-	.nomap_writes		= 0, \
-	.nomap_reads		= 0, \
-	.write_maps		= 0, \
 }
 
 struct ubd ubd_dev[MAX_DEV] = { [ 0 ... MAX_DEV - 1 ] = DEFAULT_UBD };
@@ -314,13 +297,6 @@ static int ubd_setup_common(char *str, int *index_out)
 		int major;
 
 		str++;
-		if(!strcmp(str, "mmap")){
-			CHOOSE_MODE(printk("mmap not supported by the ubd "
-					   "driver in tt mode\n"),
-				    ubd_do_mmap = 1);
-			return(0);
-		}
-
 		if(!strcmp(str, "sync")){
 			global_openflags = of_sync(global_openflags);
 			return(0);
@@ -524,7 +500,7 @@ static void ubd_handler(void)
 {
 	struct io_thread_req req;
 	struct request *rq = elv_next_request(ubd_queue);
-	int n, err;
+	int n;
 
 	do_ubd = NULL;
 	intr_count++;
@@ -538,19 +514,6 @@ static void ubd_handler(void)
 		return;
 	}
         
-	if((req.op != UBD_MMAP) &&
-	   ((req.offset != ((__u64) (rq->sector)) << 9) ||
-	    (req.length != (rq->current_nr_sectors) << 9)))
-		panic("I/O op mismatch");
-	
-	if(req.map_fd != -1){
-		err = physmem_subst_mapping(req.buffer, req.map_fd,
-					    req.map_offset, 1);
-		if(err)
-			printk("ubd_handler - physmem_subst_mapping failed, "
-			       "err = %d\n", -err);
-	}
-
 	ubd_finish(rq, req.error);
 	reactivate_fd(thread_fd, UBD_IRQ);	
 	do_ubd_request(ubd_queue);
@@ -583,14 +546,10 @@ static int ubd_file_size(struct ubd *dev, __u64 *size_out)
 
 static void ubd_close(struct ubd *dev)
 {
-	if(ubd_do_mmap)
-		physmem_forget_descriptor(dev->fd);
 	os_close_file(dev->fd);
 	if(dev->cow.file == NULL)
 		return;
 
-	if(ubd_do_mmap)
-		physmem_forget_descriptor(dev->cow.fd);
 	os_close_file(dev->cow.fd);
 	vfree(dev->cow.bitmap);
 	dev->cow.bitmap = NULL;
@@ -1010,94 +969,13 @@ static void cowify_req(struct io_thread_req *req, unsigned long *bitmap,
 			   req->bitmap_words, bitmap_len);
 }
 
-static int mmap_fd(struct request *req, struct ubd *dev, __u64 offset)
-{
-	__u64 sector;
-	unsigned char *bitmap;
-	int bit, i;
-
-	/* mmap must have been requested on the command line */
-	if(!ubd_do_mmap)
-		return(-1);
-
-	/* The buffer must be page aligned */
-	if(((unsigned long) req->buffer % UBD_MMAP_BLOCK_SIZE) != 0)
-		return(-1);
-
-	/* The request must be a page long */
-	if((req->current_nr_sectors << 9) != PAGE_SIZE)
-		return(-1);
-
-	if(dev->cow.file == NULL)
-		return(dev->fd);
-
-	sector = offset >> 9;
-	bitmap = (unsigned char *) dev->cow.bitmap;
-	bit = ubd_test_bit(sector, bitmap);
-
-	for(i = 1; i < req->current_nr_sectors; i++){
-		if(ubd_test_bit(sector + i, bitmap) != bit)
-			return(-1);
-	}
-
-	if(bit || (rq_data_dir(req) == WRITE))
-		offset += dev->cow.data_offset;
-
-	/* The data on disk must be page aligned */
-	if((offset % UBD_MMAP_BLOCK_SIZE) != 0)
-		return(-1);
-
-	return(bit ? dev->fd : dev->cow.fd);
-}
-
-static int prepare_mmap_request(struct ubd *dev, int fd, __u64 offset,
-				struct request *req,
-				struct io_thread_req *io_req)
-{
-	int err;
-
-	if(rq_data_dir(req) == WRITE){
-		/* Writes are almost no-ops since the new data is already in the
-		 * host page cache
-		 */
-		dev->map_writes++;
-		if(dev->cow.file != NULL)
-			cowify_bitmap(io_req->offset, io_req->length,
-				      &io_req->sector_mask, &io_req->cow_offset,
-				      dev->cow.bitmap, dev->cow.bitmap_offset,
-				      io_req->bitmap_words,
-				      dev->cow.bitmap_len);
-	}
-	else {
-		int w;
-
-		if((dev->cow.file != NULL) && (fd == dev->cow.fd))
-			w = 0;
-		else w = dev->openflags.w;
-
-		if((dev->cow.file != NULL) && (fd == dev->fd))
-			offset += dev->cow.data_offset;
-
-		err = physmem_subst_mapping(req->buffer, fd, offset, w);
-		if(err){
-			printk("physmem_subst_mapping failed, err = %d\n",
-			       -err);
-			return(1);
-		}
-		dev->map_reads++;
-	}
-	io_req->op = UBD_MMAP;
-	io_req->buffer = req->buffer;
-	return(0);
-}
-
 /* Called with ubd_io_lock held */
 static int prepare_request(struct request *req, struct io_thread_req *io_req)
 {
 	struct gendisk *disk = req->rq_disk;
 	struct ubd *dev = disk->private_data;
 	__u64 offset;
-	int len, fd;
+	int len;
 
 	if(req->rq_status == RQ_INACTIVE) return(1);
 
@@ -1114,34 +992,12 @@ static int prepare_request(struct request *req, struct io_thread_req *io_req)
 
 	io_req->fds[0] = (dev->cow.file != NULL) ? dev->cow.fd : dev->fd;
 	io_req->fds[1] = dev->fd;
-	io_req->map_fd = -1;
 	io_req->cow_offset = -1;
 	io_req->offset = offset;
 	io_req->length = len;
 	io_req->error = 0;
 	io_req->sector_mask = 0;
 
-	fd = mmap_fd(req, dev, io_req->offset);
-	if(fd > 0){
-		/* If mmapping is otherwise OK, but the first access to the
-		 * page is a write, then it's not mapped in yet.  So we have
-		 * to write the data to disk first, then we can map the disk
-		 * page in and continue normally from there.
-		 */
-		if((rq_data_dir(req) == WRITE) && !is_remapped(req->buffer)){
-			io_req->map_fd = dev->fd;
-			io_req->map_offset = io_req->offset +
-				dev->cow.data_offset;
-			dev->write_maps++;
-		}
-		else return(prepare_mmap_request(dev, fd, io_req->offset, req,
-						 io_req));
-	}
-
-	if(rq_data_dir(req) == READ)
-		dev->nomap_reads++;
-	else dev->nomap_writes++;
-
 	io_req->op = (rq_data_dir(req) == READ) ? UBD_READ : UBD_WRITE;
 	io_req->offsets[0] = 0;
 	io_req->offsets[1] = dev->cow.data_offset;
@@ -1229,143 +1085,6 @@ static int ubd_ioctl(struct inode * inode, struct file * file,
 	return(-EINVAL);
 }
 
-static int ubd_check_remapped(int fd, unsigned long address, int is_write,
-			      __u64 offset)
-{
-	__u64 bitmap_offset;
-	unsigned long new_bitmap[2];
-	int i, err, n;
-
-	/* If it's not a write access, we can't do anything about it */
-	if(!is_write)
-		return(0);
-
-	/* We have a write */
-	for(i = 0; i < sizeof(ubd_dev) / sizeof(ubd_dev[0]); i++){
-		struct ubd *dev = &ubd_dev[i];
-
-		if((dev->fd != fd) && (dev->cow.fd != fd))
-			continue;
-
-		/* It's a write to a ubd device */
-
-		/* This should be impossible now */
-		if(!dev->openflags.w){
-			/* It's a write access on a read-only device - probably
-			 * shouldn't happen.  If the kernel is trying to change
-			 * something with no intention of writing it back out,
-			 * then this message will clue us in that this needs
-			 * fixing
-			 */
-			printk("Write access to mapped page from readonly ubd "
-			       "device %d\n", i);
-			return(0);
-		}
-
-		/* It's a write to a writeable ubd device - it must be COWed
-		 * because, otherwise, the page would have been mapped in
-		 * writeable
-		 */
-
-		if(!dev->cow.file)
-			panic("Write fault on writeable non-COW ubd device %d",
-			      i);
-
-		/* It should also be an access to the backing file since the
-		 * COW pages should be mapped in read-write
-		 */
-
-		if(fd == dev->fd)
-			panic("Write fault on a backing page of ubd "
-			      "device %d\n", i);
-
-		/* So, we do the write, copying the backing data to the COW
-		 * file...
-		 */
-
-		err = os_seek_file(dev->fd, offset + dev->cow.data_offset);
-		if(err < 0)
-			panic("Couldn't seek to %lld in COW file of ubd "
-			      "device %d, err = %d",
-			      offset + dev->cow.data_offset, i, -err);
-
-		n = os_write_file(dev->fd, (void *) address, PAGE_SIZE);
-		if(n != PAGE_SIZE)
-			panic("Couldn't copy data to COW file of ubd "
-			      "device %d, err = %d", i, -n);
-
-		/* ... updating the COW bitmap... */
-
-		cowify_bitmap(offset, PAGE_SIZE, NULL, &bitmap_offset,
-			      dev->cow.bitmap, dev->cow.bitmap_offset,
-			      new_bitmap, dev->cow.bitmap_len);
-
-		err = os_seek_file(dev->fd, bitmap_offset);
-		if(err < 0)
-			panic("Couldn't seek to %lld in COW file of ubd "
-			      "device %d, err = %d", bitmap_offset, i, -err);
-
-		n = os_write_file(dev->fd, new_bitmap, sizeof(new_bitmap));
-		if(n != sizeof(new_bitmap))
-			panic("Couldn't update bitmap  of ubd device %d, "
-			      "err = %d", i, -n);
-
-		/* Maybe we can map the COW page in, and maybe we can't.  If
-		 * it is a pre-V3 COW file, we can't, since the alignment will
-		 * be wrong.  If it is a V3 or later COW file which has been
-		 * moved to a system with a larger page size, then maybe we
-		 * can't, depending on the exact location of the page.
-		 */
-
-		offset += dev->cow.data_offset;
-
-		/* Remove the remapping, putting the original anonymous page
-		 * back.  If the COW file can be mapped in, that is done.
-		 * Otherwise, the COW page is read in.
-		 */
-
-		if(!physmem_remove_mapping((void *) address))
-			panic("Address 0x%lx not remapped by ubd device %d",
-			      address, i);
-		if((offset % UBD_MMAP_BLOCK_SIZE) == 0)
-			physmem_subst_mapping((void *) address, dev->fd,
-					      offset, 1);
-		else {
-			err = os_seek_file(dev->fd, offset);
-			if(err < 0)
-				panic("Couldn't seek to %lld in COW file of "
-				      "ubd device %d, err = %d", offset, i,
-				      -err);
-
-			n = os_read_file(dev->fd, (void *) address, PAGE_SIZE);
-			if(n != PAGE_SIZE)
-				panic("Failed to read page from offset %llx of "
-				      "COW file of ubd device %d, err = %d",
-				      offset, i, -n);
-		}
-
-		return(1);
-	}
-
-	/* It's not a write on a ubd device */
-	return(0);
-}
-
-static struct remapper ubd_remapper = {
-	.list	= LIST_HEAD_INIT(ubd_remapper.list),
-	.proc	= ubd_check_remapped,
-};
-
-static int ubd_remapper_setup(void)
-{
-	if(ubd_do_mmap)
-		register_remapper(&ubd_remapper);
-
-	return(0);
-}
-
-__initcall(ubd_remapper_setup);
-
 static int same_backing_files(char *from_cmdline, char *from_cow, char *cow)
 {
 	struct uml_stat buf1, buf2;
@@ -1568,15 +1287,6 @@ void do_io(struct io_thread_req *req)
 	int err;
 	__u64 off;
 
-	if(req->op == UBD_MMAP){
-		/* Touch the page to force the host to do any necessary IO to
-		 * get it into memory
-		 */
-		n = *((volatile int *) req->buffer);
-		req->error = update_bitmap(req);
-		return;
-	}
-
 	nsectors = req->length / req->sectorsize;
 	start = 0;
 	do {
diff --git a/arch/um/include/sysdep-i386/ptrace.h b/arch/um/include/sysdep-i386/ptrace.h
index 84ec7ff5cf8..6eaeb991998 100644
--- a/arch/um/include/sysdep-i386/ptrace.h
+++ b/arch/um/include/sysdep-i386/ptrace.h
@@ -31,7 +31,6 @@ extern int sysemu_supported;
 #ifdef UML_CONFIG_MODE_SKAS
 
 #include "skas_ptregs.h"
-#include "sysdep/faultinfo.h"
 
 #define REGS_IP(r) ((r)[HOST_IP])
 #define REGS_SP(r) ((r)[HOST_SP])
@@ -59,6 +58,7 @@ extern int sysemu_supported;
 #define PTRACE_SYSEMU_SINGLESTEP 32
 #endif
 
+#include "sysdep/faultinfo.h"
 #include "choose-mode.h"
 
 union uml_pt_regs {
diff --git a/arch/um/include/sysdep-x86_64/checksum.h b/arch/um/include/sysdep-x86_64/checksum.h
index 572c6c19be3..ea97005af69 100644
--- a/arch/um/include/sysdep-x86_64/checksum.h
+++ b/arch/um/include/sysdep-x86_64/checksum.h
@@ -9,8 +9,6 @@
 #include "linux/in6.h"
 #include "asm/uaccess.h"
 
-extern unsigned int csum_partial_copy_from(const unsigned char *src, unsigned char *dst, int len,
-					   int sum, int *err_ptr);
 extern unsigned csum_partial(const unsigned char *buff, unsigned len,
                              unsigned sum);
 
@@ -31,10 +29,15 @@ unsigned int csum_partial_copy_nocheck(const unsigned char *src, unsigned char *
 }
 
 static __inline__
-unsigned int csum_partial_copy_from_user(const unsigned char *src, unsigned char *dst,
-					 int len, int sum, int *err_ptr)
+unsigned int csum_partial_copy_from_user(const unsigned char *src,
+                                         unsigned char *dst, int len, int sum,
+                                         int *err_ptr)
 {
-	return csum_partial_copy_from(src, dst, len, sum, err_ptr);
+        if(copy_from_user(dst, src, len)){
+                *err_ptr = -EFAULT;
+                return(-1);
+        }
+        return csum_partial(dst, len, sum);
 }
 
 /**
@@ -137,15 +140,6 @@ static inline unsigned add32_with_carry(unsigned a, unsigned b)
         return a;
 }
 
-#endif
+extern unsigned short ip_compute_csum(unsigned char * buff, int len);
 
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
+#endif
diff --git a/arch/um/include/sysdep-x86_64/ptrace.h b/arch/um/include/sysdep-x86_64/ptrace.h
index 348e8fcd513..be8acd5efd9 100644
--- a/arch/um/include/sysdep-x86_64/ptrace.h
+++ b/arch/um/include/sysdep-x86_64/ptrace.h
@@ -135,6 +135,7 @@ extern int mode_tt;
 	__CHOOSE_MODE(SC_EFLAGS(UPT_SC(r)), REGS_EFLAGS((r)->skas.regs))
 #define UPT_SC(r) ((r)->tt.sc)
 #define UPT_SYSCALL_NR(r) __CHOOSE_MODE((r)->tt.syscall, (r)->skas.syscall)
+#define UPT_SYSCALL_RET(r) UPT_RAX(r)
 
 extern int user_context(unsigned long sp);
 
@@ -196,32 +197,32 @@ struct syscall_args {
 
 
 #define UPT_SET(regs, reg, val) \
-        ({      unsigned long val; \
+        ({      unsigned long __upt_val = val; \
                 switch(reg){ \
-		case R8: UPT_R8(regs) = val; break; \
-		case R9: UPT_R9(regs) = val; break; \
-		case R10: UPT_R10(regs) = val; break; \
-		case R11: UPT_R11(regs) = val; break; \
-		case R12: UPT_R12(regs) = val; break; \
-		case R13: UPT_R13(regs) = val; break; \
-		case R14: UPT_R14(regs) = val; break; \
-		case R15: UPT_R15(regs) = val; break; \
-                case RIP: UPT_IP(regs) = val; break; \
-                case RSP: UPT_SP(regs) = val; break; \
-                case RAX: UPT_RAX(regs) = val; break; \
-                case RBX: UPT_RBX(regs) = val; break; \
-                case RCX: UPT_RCX(regs) = val; break; \
-                case RDX: UPT_RDX(regs) = val; break; \
-                case RSI: UPT_RSI(regs) = val; break; \
-                case RDI: UPT_RDI(regs) = val; break; \
-                case RBP: UPT_RBP(regs) = val; break; \
-                case ORIG_RAX: UPT_ORIG_RAX(regs) = val; break; \
-                case CS: UPT_CS(regs) = val; break; \
-                case DS: UPT_DS(regs) = val; break; \
-                case ES: UPT_ES(regs) = val; break; \
-                case FS: UPT_FS(regs) = val; break; \
-                case GS: UPT_GS(regs) = val; break; \
-                case EFLAGS: UPT_EFLAGS(regs) = val; break; \
+                case R8: UPT_R8(regs) = __upt_val; break; \
+                case R9: UPT_R9(regs) = __upt_val; break; \
+                case R10: UPT_R10(regs) = __upt_val; break; \
+                case R11: UPT_R11(regs) = __upt_val; break; \
+                case R12: UPT_R12(regs) = __upt_val; break; \
+                case R13: UPT_R13(regs) = __upt_val; break; \
+                case R14: UPT_R14(regs) = __upt_val; break; \
+                case R15: UPT_R15(regs) = __upt_val; break; \
+                case RIP: UPT_IP(regs) = __upt_val; break; \
+                case RSP: UPT_SP(regs) = __upt_val; break; \
+                case RAX: UPT_RAX(regs) = __upt_val; break; \
+                case RBX: UPT_RBX(regs) = __upt_val; break; \
+                case RCX: UPT_RCX(regs) = __upt_val; break; \
+                case RDX: UPT_RDX(regs) = __upt_val; break; \
+                case RSI: UPT_RSI(regs) = __upt_val; break; \
+                case RDI: UPT_RDI(regs) = __upt_val; break; \
+                case RBP: UPT_RBP(regs) = __upt_val; break; \
+                case ORIG_RAX: UPT_ORIG_RAX(regs) = __upt_val; break; \
+                case CS: UPT_CS(regs) = __upt_val; break; \
+                case DS: UPT_DS(regs) = __upt_val; break; \
+                case ES: UPT_ES(regs) = __upt_val; break; \
+                case FS: UPT_FS(regs) = __upt_val; break; \
+                case GS: UPT_GS(regs) = __upt_val; break; \
+                case EFLAGS: UPT_EFLAGS(regs) = __upt_val; break; \
                 default :  \
                         panic("Bad register in UPT_SET : %d\n", reg);  \
 			break; \
@@ -245,14 +246,3 @@ struct syscall_args {
         CHOOSE_MODE((&(r)->tt.faultinfo), (&(r)->skas.faultinfo))
 
 #endif
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
diff --git a/arch/um/kernel/Makefile b/arch/um/kernel/Makefile
index 9736ca27c5f..a8918e80df9 100644
--- a/arch/um/kernel/Makefile
+++ b/arch/um/kernel/Makefile
@@ -14,7 +14,7 @@ obj-y = config.o exec_kern.o exitcode.o \
 	tlb.o trap_kern.o trap_user.o uaccess_user.o um_arch.o umid.o \
 	user_util.o
 
-obj-$(CONFIG_BLK_DEV_INITRD) += initrd_kern.o initrd_user.o
+obj-$(CONFIG_BLK_DEV_INITRD) += initrd.o
 obj-$(CONFIG_GPROF)	+= gprof_syms.o
 obj-$(CONFIG_GCOV)	+= gmon_syms.o
 obj-$(CONFIG_TTY_LOG)	+= tty_log.o
diff --git a/arch/um/kernel/checksum.c b/arch/um/kernel/checksum.c
deleted file mode 100644
index e69de29bb2d..00000000000
--- a/arch/um/kernel/checksum.c
+++ /dev/null
diff --git a/arch/um/kernel/initrd.c b/arch/um/kernel/initrd.c
new file mode 100644
index 00000000000..82ecf904b09
--- /dev/null
+++ b/arch/um/kernel/initrd.c
@@ -0,0 +1,78 @@
+/*
+ * Copyright (C) 2000, 2001, 2002 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include "linux/init.h"
+#include "linux/bootmem.h"
+#include "linux/initrd.h"
+#include "asm/types.h"
+#include "user_util.h"
+#include "kern_util.h"
+#include "initrd.h"
+#include "init.h"
+#include "os.h"
+
+/* Changed by uml_initrd_setup, which is a setup */
+static char *initrd __initdata = NULL;
+
+static int __init read_initrd(void)
+{
+	void *area;
+	long long size;
+	int err;
+
+	if(initrd == NULL) return 0;
+	err = os_file_size(initrd, &size);
+	if(err) return 0;
+	area = alloc_bootmem(size);
+	if(area == NULL) return 0;
+	if(load_initrd(initrd, area, size) == -1) return 0;
+	initrd_start = (unsigned long) area;
+	initrd_end = initrd_start + size;
+	return 0;
+}
+
+__uml_postsetup(read_initrd);
+
+static int __init uml_initrd_setup(char *line, int *add)
+{
+	initrd = line;
+	return 0;
+}
+
+__uml_setup("initrd=", uml_initrd_setup,
+"initrd=<initrd image>\n"
+"    This is used to boot UML from an initrd image.  The argument is the\n"
+"    name of the file containing the image.\n\n"
+);
+
+int load_initrd(char *filename, void *buf, int size)
+{
+	int fd, n;
+
+	fd = os_open_file(filename, of_read(OPENFLAGS()), 0);
+	if(fd < 0){
+		printk("Opening '%s' failed - err = %d\n", filename, -fd);
+		return(-1);
+	}
+	n = os_read_file(fd, buf, size);
+	if(n != size){
+		printk("Read of %d bytes from '%s' failed, err = %d\n", size,
+		       filename, -n);
+		return(-1);
+	}
+
+	os_close_file(fd);
+	return(0);
+}
+/*
+ * Overrides for Emacs so that we follow Linus's tabbing style.
+ * Emacs will notice this stuff at the end of the file and automatically
+ * adjust the settings for this buffer only.  This must remain at the end
+ * of the file.
+ * ---------------------------------------------------------------------------
+ * Local variables:
+ * c-file-style: "linux"
+ * End:
+ */
diff --git a/arch/um/kernel/irq_user.c b/arch/um/kernel/irq_user.c
index 6d6f9484b88..b3074cbaa47 100644
--- a/arch/um/kernel/irq_user.c
+++ b/arch/um/kernel/irq_user.c
@@ -236,9 +236,15 @@ static void free_irq_by_cb(int (*test)(struct irq_fd *, void *), void *arg)
 				       (*prev)->fd, pollfds[i].fd);
 				goto out;
 			}
-			memcpy(&pollfds[i], &pollfds[i + 1],
-			       (pollfds_num - i - 1) * sizeof(pollfds[0]));
+
 			pollfds_num--;
+
+			/* This moves the *whole* array after pollfds[i] (though
+			 * it doesn't spot as such)! */
+
+			memmove(&pollfds[i], &pollfds[i + 1],
+			       (pollfds_num - i) * sizeof(pollfds[0]));
+
 			if(last_irq_ptr == &old_fd->next) 
 				last_irq_ptr = prev;
 			*prev = (*prev)->next;
diff --git a/arch/um/kernel/ksyms.c b/arch/um/kernel/ksyms.c
index 78d69dc74b2..99439fa15ef 100644
--- a/arch/um/kernel/ksyms.c
+++ b/arch/um/kernel/ksyms.c
@@ -57,6 +57,7 @@ EXPORT_SYMBOL(copy_to_user_tt);
 EXPORT_SYMBOL(strncpy_from_user_skas);
 EXPORT_SYMBOL(copy_to_user_skas);
 EXPORT_SYMBOL(copy_from_user_skas);
+EXPORT_SYMBOL(clear_user_skas);
 #endif
 EXPORT_SYMBOL(uml_strdup);
 
diff --git a/arch/um/kernel/mem.c b/arch/um/kernel/mem.c
index f156661781c..c22825f13e4 100644
--- a/arch/um/kernel/mem.c
+++ b/arch/um/kernel/mem.c
@@ -100,12 +100,37 @@ void mem_init(void)
 #endif
 }
 
+/*
+ * Create a page table and place a pointer to it in a middle page
+ * directory entry.
+ */
+static void __init one_page_table_init(pmd_t *pmd)
+{
+	if (pmd_none(*pmd)) {
+		pte_t *pte = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+		set_pmd(pmd, __pmd(_KERNPG_TABLE +
+					   (unsigned long) __pa(pte)));
+		if (pte != pte_offset_kernel(pmd, 0))
+			BUG();
+	}
+}
+
+static void __init one_md_table_init(pud_t *pud)
+{
+#ifdef CONFIG_3_LEVEL_PGTABLES
+	pmd_t *pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+	set_pud(pud, __pud(_KERNPG_TABLE + (unsigned long) __pa(pmd_table)));
+	if (pmd_table != pmd_offset(pud, 0))
+		BUG();
+#endif
+}
+
 static void __init fixrange_init(unsigned long start, unsigned long end, 
 				 pgd_t *pgd_base)
 {
 	pgd_t *pgd;
+	pud_t *pud;
 	pmd_t *pmd;
-	pte_t *pte;
 	int i, j;
 	unsigned long vaddr;
 
@@ -115,15 +140,12 @@ static void __init fixrange_init(unsigned long start, unsigned long end,
 	pgd = pgd_base + i;
 
 	for ( ; (i < PTRS_PER_PGD) && (vaddr < end); pgd++, i++) {
-		pmd = (pmd_t *)pgd;
+		pud = pud_offset(pgd, vaddr);
+		if (pud_none(*pud))
+			one_md_table_init(pud);
+		pmd = pmd_offset(pud, vaddr);
 		for (; (j < PTRS_PER_PMD) && (vaddr != end); pmd++, j++) {
-			if (pmd_none(*pmd)) {
-				pte = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE);
-				set_pmd(pmd, __pmd(_KERNPG_TABLE + 
-						   (unsigned long) __pa(pte)));
-				if (pte != pte_offset_kernel(pmd, 0))
-					BUG();
-			}
+			one_page_table_init(pmd);
 			vaddr += PMD_SIZE;
 		}
 		j = 0;
diff --git a/arch/um/kernel/ptrace.c b/arch/um/kernel/ptrace.c
index 2b75d8d9ba7..2925e15324d 100644
--- a/arch/um/kernel/ptrace.c
+++ b/arch/um/kernel/ptrace.c
@@ -28,9 +28,9 @@ static inline void set_singlestepping(struct task_struct *child, int on)
         child->thread.singlestep_syscall = 0;
 
 #ifdef SUBARCH_SET_SINGLESTEPPING
-        SUBARCH_SET_SINGLESTEPPING(child, on)
+        SUBARCH_SET_SINGLESTEPPING(child, on);
 #endif
-                }
+}
 
 /*
  * Called by kernel/ptrace.c when detaching..
@@ -83,7 +83,7 @@ long sys_ptrace(long request, long pid, long addr, long data)
 	}
 
 #ifdef SUBACH_PTRACE_SPECIAL
-        SUBARCH_PTRACE_SPECIAL(child,request,addr,data)
+        SUBARCH_PTRACE_SPECIAL(child,request,addr,data);
 #endif
 
 	ret = ptrace_check_attach(child, request == PTRACE_KILL);
diff --git a/arch/um/kernel/trap_kern.c b/arch/um/kernel/trap_kern.c
index 5fca2c61eb9..1de22d8a313 100644
--- a/arch/um/kernel/trap_kern.c
+++ b/arch/um/kernel/trap_kern.c
@@ -57,10 +57,11 @@ int handle_page_fault(unsigned long address, unsigned long ip,
 	*code_out = SEGV_ACCERR;
 	if(is_write && !(vma->vm_flags & VM_WRITE)) 
 		goto out;
+
+        if(!(vma->vm_flags & (VM_READ | VM_EXEC)))
+                goto out;
+
 	page = address & PAGE_MASK;
-	pgd = pgd_offset(mm, page);
-	pud = pud_offset(pgd, page);
-	pmd = pmd_offset(pud, page);
 	do {
  survive:
 		switch (handle_mm_fault(mm, vma, address, is_write)){
@@ -106,33 +107,6 @@ out_of_memory:
 	goto out;
 }
 
-LIST_HEAD(physmem_remappers);
-
-void register_remapper(struct remapper *info)
-{
-	list_add(&info->list, &physmem_remappers);
-}
-
-static int check_remapped_addr(unsigned long address, int is_write)
-{
-	struct remapper *remapper;
-	struct list_head *ele;
-	__u64 offset;
-	int fd;
-
-	fd = phys_mapping(__pa(address), &offset);
-	if(fd == -1)
-		return(0);
-
-	list_for_each(ele, &physmem_remappers){
-		remapper = list_entry(ele, struct remapper, list);
-		if((*remapper->proc)(fd, address, is_write, offset))
-			return(1);
-	}
-
-	return(0);
-}
-
 /*
  * We give a *copy* of the faultinfo in the regs to segv.
  * This must be done, since nesting SEGVs could overwrite
@@ -151,8 +125,6 @@ unsigned long segv(struct faultinfo fi, unsigned long ip, int is_user, void *sc)
                 flush_tlb_kernel_vm();
                 return(0);
         }
-	else if(check_remapped_addr(address & PAGE_MASK, is_write))
-		return(0);
 	else if(current->mm == NULL)
 		panic("Segfault with no mm");
 	err = handle_page_fault(address, ip, is_write, is_user, &si.si_code);
diff --git a/arch/um/kernel/tt/ksyms.c b/arch/um/kernel/tt/ksyms.c
index 92ec85d67c7..84a9385a8fe 100644
--- a/arch/um/kernel/tt/ksyms.c
+++ b/arch/um/kernel/tt/ksyms.c
@@ -12,6 +12,7 @@ EXPORT_SYMBOL(__do_copy_to_user);
 EXPORT_SYMBOL(__do_strncpy_from_user);
 EXPORT_SYMBOL(__do_strnlen_user); 
 EXPORT_SYMBOL(__do_clear_user);
+EXPORT_SYMBOL(clear_user_tt);
 
 EXPORT_SYMBOL(tracing_pid);
 EXPORT_SYMBOL(honeypot);
diff --git a/arch/um/kernel/uml.lds.S b/arch/um/kernel/uml.lds.S
index 76eadb30918..dd5355500bd 100644
--- a/arch/um/kernel/uml.lds.S
+++ b/arch/um/kernel/uml.lds.S
@@ -73,6 +73,8 @@ SECTIONS
 
   .got           : { *(.got.plt) *(.got) }
   .dynamic       : { *(.dynamic) }
+  .tdata	  : { *(.tdata .tdata.* .gnu.linkonce.td.*) }
+  .tbss		  : { *(.tbss .tbss.* .gnu.linkonce.tb.*) *(.tcommon) }
   /* We want the small data sections together, so single-instruction offsets
      can access them all, and initialized data all before uninitialized, so
      we can shorten the on-disk segment size.  */
diff --git a/arch/um/sys-i386/Makefile b/arch/um/sys-i386/Makefile
index fcd67c3414e..4351e560550 100644
--- a/arch/um/sys-i386/Makefile
+++ b/arch/um/sys-i386/Makefile
@@ -9,11 +9,11 @@ USER_OBJS := bugs.o ptrace_user.o sigcontext.o fault.o
 
 SYMLINKS = bitops.c semaphore.c highmem.c module.c
 
+include arch/um/scripts/Makefile.rules
+
 bitops.c-dir = lib
 semaphore.c-dir = kernel
 highmem.c-dir = mm
 module.c-dir = kernel
 
 subdir- := util
-
-include arch/um/scripts/Makefile.rules
diff --git a/arch/um/sys-i386/delay.c b/arch/um/sys-i386/delay.c
index e9892eef51c..2c11b9770e8 100644
--- a/arch/um/sys-i386/delay.c
+++ b/arch/um/sys-i386/delay.c
@@ -1,5 +1,7 @@
-#include "linux/delay.h"
-#include "asm/param.h"
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/delay.h>
+#include <asm/param.h>
 
 void __delay(unsigned long time)
 {
@@ -20,13 +22,19 @@ void __udelay(unsigned long usecs)
 	int i, n;
 
 	n = (loops_per_jiffy * HZ * usecs) / MILLION;
-	for(i=0;i<n;i++) ;
+        for(i=0;i<n;i++)
+                cpu_relax();
 }
 
+EXPORT_SYMBOL(__udelay);
+
 void __const_udelay(unsigned long usecs)
 {
 	int i, n;
 
 	n = (loops_per_jiffy * HZ * usecs) / MILLION;
-	for(i=0;i<n;i++) ;
+        for(i=0;i<n;i++)
+                cpu_relax();
 }
+
+EXPORT_SYMBOL(__const_udelay);
diff --git a/arch/um/sys-x86_64/Makefile b/arch/um/sys-x86_64/Makefile
index 3d7da911cc8..608466ad6b2 100644
--- a/arch/um/sys-x86_64/Makefile
+++ b/arch/um/sys-x86_64/Makefile
@@ -14,11 +14,11 @@ obj-$(CONFIG_MODULES) += module.o um_module.o
 
 USER_OBJS := ptrace_user.o sigcontext.o
 
-include arch/um/scripts/Makefile.rules
-
 SYMLINKS = bitops.c csum-copy.S csum-partial.c csum-wrappers.c memcpy.S \
 	semaphore.c thunk.S module.c
 
+include arch/um/scripts/Makefile.rules
+
 bitops.c-dir = lib
 csum-copy.S-dir = lib
 csum-partial.c-dir = lib
@@ -28,6 +28,4 @@ semaphore.c-dir = kernel
 thunk.S-dir = lib
 module.c-dir = kernel
 
-CFLAGS_csum-partial.o := -Dcsum_partial=arch_csum_partial
-
 subdir- := util
diff --git a/arch/um/sys-x86_64/delay.c b/arch/um/sys-x86_64/delay.c
index 651332aeec2..137f4446b43 100644
--- a/arch/um/sys-x86_64/delay.c
+++ b/arch/um/sys-x86_64/delay.c
@@ -5,40 +5,37 @@
  * Licensed under the GPL
  */
 
-#include "linux/delay.h"
-#include "asm/processor.h"
-#include "asm/param.h"
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <asm/processor.h>
+#include <asm/param.h>
 
 void __delay(unsigned long loops)
 {
 	unsigned long i;
 
-	for(i = 0; i < loops; i++) ;
+        for(i = 0; i < loops; i++)
+                cpu_relax();
 }
 
 void __udelay(unsigned long usecs)
 {
-	int i, n;
+	unsigned long i, n;
 
 	n = (loops_per_jiffy * HZ * usecs) / MILLION;
-	for(i=0;i<n;i++) ;
+        for(i=0;i<n;i++)
+                cpu_relax();
 }
 
+EXPORT_SYMBOL(__udelay);
+
 void __const_udelay(unsigned long usecs)
 {
-	int i, n;
+	unsigned long i, n;
 
 	n = (loops_per_jiffy * HZ * usecs) / MILLION;
-	for(i=0;i<n;i++) ;
+        for(i=0;i<n;i++)
+                cpu_relax();
 }
 
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
+EXPORT_SYMBOL(__const_udelay);
diff --git a/arch/um/sys-x86_64/ksyms.c b/arch/um/sys-x86_64/ksyms.c
index a27f0ee6a4f..85927380820 100644
--- a/arch/um/sys-x86_64/ksyms.c
+++ b/arch/um/sys-x86_64/ksyms.c
@@ -16,5 +16,4 @@ EXPORT_SYMBOL(__up_wakeup);
 EXPORT_SYMBOL(__memcpy);
 
 /* Networking helper routines. */
-/*EXPORT_SYMBOL(csum_partial_copy_from);
-EXPORT_SYMBOL(csum_partial_copy_to);*/
+EXPORT_SYMBOL(ip_compute_csum);
diff --git a/arch/um/sys-x86_64/ptrace.c b/arch/um/sys-x86_64/ptrace.c
index b593bb256f2..74eee5c7c6d 100644
--- a/arch/um/sys-x86_64/ptrace.c
+++ b/arch/um/sys-x86_64/ptrace.c
@@ -5,10 +5,11 @@
  */
 
 #define __FRAME_OFFSETS
-#include "asm/ptrace.h"
-#include "linux/sched.h"
-#include "linux/errno.h"
-#include "asm/elf.h"
+#include <asm/ptrace.h>
+#include <linux/sched.h>
+#include <linux/errno.h>
+#include <asm/uaccess.h>
+#include <asm/elf.h>
 
 /* XXX x86_64 */
 unsigned long not_ss;
diff --git a/arch/um/sys-x86_64/syscalls.c b/arch/um/sys-x86_64/syscalls.c
index dd9914642b8..d4a59657fb9 100644
--- a/arch/um/sys-x86_64/syscalls.c
+++ b/arch/um/sys-x86_64/syscalls.c
@@ -15,6 +15,7 @@
 #include "asm/unistd.h"
 #include "asm/prctl.h" /* XXX This should get the constants from libc */
 #include "choose-mode.h"
+#include "kern.h"
 
 asmlinkage long sys_uname64(struct new_utsname __user * name)
 {
diff --git a/arch/um/sys-x86_64/user-offsets.c b/arch/um/sys-x86_64/user-offsets.c
index 5e14792e483..513d17ceafd 100644
--- a/arch/um/sys-x86_64/user-offsets.c
+++ b/arch/um/sys-x86_64/user-offsets.c
@@ -3,6 +3,14 @@
 #include <signal.h>
 #define __FRAME_OFFSETS
 #include <asm/ptrace.h>
+#include <asm/types.h>
+/* For some reason, x86_64 defines u64 and u32 only in <pci/types.h>, which I
+ * refuse to include here, even though they're used throughout the headers.
+ * These are used in asm/user.h, and that include can't be avoided because of
+ * the sizeof(struct user_regs_struct) below.
+ */
+typedef __u64 u64;
+typedef __u32 u32;
 #include <asm/user.h>
 
 #define DEFINE(sym, val) \
diff --git a/arch/x86_64/Kconfig b/arch/x86_64/Kconfig
index 44ee7f6acf7..82cb2a3f127 100644
--- a/arch/x86_64/Kconfig
+++ b/arch/x86_64/Kconfig
@@ -303,6 +303,20 @@ config HPET_TIMER
 	  as it is off-chip.  You can find the HPET spec at
 	  <http://www.intel.com/labs/platcomp/hpet/hpetspec.htm>.
 
+config X86_PM_TIMER
+	bool "PM timer"
+	default y
+	help
+	  Support the ACPI PM timer for time keeping. This is slow,
+	  but is useful on some chipsets without HPET on systems with more
+	  than one CPU. On a single processor or single socket multi core
+	  system it is normally not required.
+	  When the PM timer is active 64bit vsyscalls are disabled
+	  and should not be enabled (/proc/sys/kernel/vsyscall64 should
+	  not be changed).
+	  The kernel selects the PM timer only as a last resort, so it is
+	  useful to enable just in case.
+
 config HPET_EMULATE_RTC
 	bool "Provide RTC interrupt"
 	depends on HPET_TIMER && RTC=y
diff --git a/arch/x86_64/defconfig b/arch/x86_64/defconfig
index 9ce51dee30b..569595b74c7 100644
--- a/arch/x86_64/defconfig
+++ b/arch/x86_64/defconfig
@@ -1,7 +1,7 @@
 #
 # Automatically generated make config: don't edit
-# Linux kernel version: 2.6.11-bk7
-# Sat Mar 12 23:43:44 2005
+# Linux kernel version: 2.6.12-rc4
+# Fri May 13 06:39:11 2005
 #
 CONFIG_X86_64=y
 CONFIG_64BIT=y
@@ -11,8 +11,6 @@ CONFIG_RWSEM_GENERIC_SPINLOCK=y
 CONFIG_GENERIC_CALIBRATE_DELAY=y
 CONFIG_X86_CMPXCHG=y
 CONFIG_EARLY_PRINTK=y
-CONFIG_HPET_TIMER=y
-CONFIG_HPET_EMULATE_RTC=y
 CONFIG_GENERIC_ISA_DMA=y
 CONFIG_GENERIC_IOMAP=y
 
@@ -22,6 +20,7 @@ CONFIG_GENERIC_IOMAP=y
 CONFIG_EXPERIMENTAL=y
 CONFIG_CLEAN_COMPILE=y
 CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
 
 #
 # General setup
@@ -33,7 +32,6 @@ CONFIG_POSIX_MQUEUE=y
 # CONFIG_BSD_PROCESS_ACCT is not set
 CONFIG_SYSCTL=y
 # CONFIG_AUDIT is not set
-CONFIG_LOG_BUF_SHIFT=18
 # CONFIG_HOTPLUG is not set
 CONFIG_KOBJECT_UEVENT=y
 CONFIG_IKCONFIG=y
@@ -43,10 +41,11 @@ CONFIG_IKCONFIG_PROC=y
 CONFIG_KALLSYMS=y
 CONFIG_KALLSYMS_ALL=y
 # CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_PRINTK=y
+CONFIG_BUG=y
 CONFIG_BASE_FULL=y
 CONFIG_FUTEX=y
 CONFIG_EPOLL=y
-# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
 CONFIG_SHMEM=y
 CONFIG_CC_ALIGN_FUNCTIONS=0
 CONFIG_CC_ALIGN_LABELS=0
@@ -93,6 +92,9 @@ CONFIG_DISCONTIGMEM=y
 CONFIG_NUMA=y
 CONFIG_HAVE_DEC_LOCK=y
 CONFIG_NR_CPUS=8
+CONFIG_HPET_TIMER=y
+CONFIG_X86_PM_TIMER=y
+CONFIG_HPET_EMULATE_RTC=y
 CONFIG_GART_IOMMU=y
 CONFIG_SWIOTLB=y
 CONFIG_X86_MCE=y
@@ -100,6 +102,7 @@ CONFIG_X86_MCE_INTEL=y
 CONFIG_SECCOMP=y
 CONFIG_GENERIC_HARDIRQS=y
 CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_ISA_DMA_API=y
 
 #
 # Power management options
@@ -129,7 +132,7 @@ CONFIG_ACPI_NUMA=y
 # CONFIG_ACPI_IBM is not set
 CONFIG_ACPI_TOSHIBA=y
 CONFIG_ACPI_BLACKLIST_YEAR=2001
-CONFIG_ACPI_DEBUG=y
+# CONFIG_ACPI_DEBUG is not set
 CONFIG_ACPI_BUS=y
 CONFIG_ACPI_EC=y
 CONFIG_ACPI_POWER=y
@@ -141,6 +144,7 @@ CONFIG_ACPI_SYSTEM=y
 # CPU Frequency scaling
 #
 CONFIG_CPU_FREQ=y
+CONFIG_CPU_FREQ_TABLE=y
 # CONFIG_CPU_FREQ_DEBUG is not set
 CONFIG_CPU_FREQ_STAT=y
 # CONFIG_CPU_FREQ_STAT_DETAILS is not set
@@ -150,7 +154,6 @@ CONFIG_CPU_FREQ_GOV_PERFORMANCE=y
 # CONFIG_CPU_FREQ_GOV_POWERSAVE is not set
 CONFIG_CPU_FREQ_GOV_USERSPACE=y
 CONFIG_CPU_FREQ_GOV_ONDEMAND=y
-CONFIG_CPU_FREQ_TABLE=y
 
 #
 # CPUFreq processor drivers
@@ -164,6 +167,7 @@ CONFIG_X86_ACPI_CPUFREQ=y
 # shared options
 #
 CONFIG_X86_ACPI_CPUFREQ_PROC_INTF=y
+# CONFIG_X86_SPEEDSTEP_LIB is not set
 
 #
 # Bus options (PCI etc.)
@@ -172,9 +176,11 @@ CONFIG_PCI=y
 CONFIG_PCI_DIRECT=y
 CONFIG_PCI_MMCONFIG=y
 CONFIG_UNORDERED_IO=y
+# CONFIG_PCIEPORTBUS is not set
 CONFIG_PCI_MSI=y
 # CONFIG_PCI_LEGACY_PROC is not set
 # CONFIG_PCI_NAMES is not set
+# CONFIG_PCI_DEBUG is not set
 
 #
 # PCCARD (PCMCIA/CardBus) support
@@ -182,10 +188,6 @@ CONFIG_PCI_MSI=y
 # CONFIG_PCCARD is not set
 
 #
-# PC-card bridges
-#
-
-#
 # PCI Hotplug Support
 #
 # CONFIG_HOTPLUG_PCI is not set
@@ -254,7 +256,7 @@ CONFIG_LBD=y
 # IO Schedulers
 #
 CONFIG_IOSCHED_NOOP=y
-CONFIG_IOSCHED_AS=y
+# CONFIG_IOSCHED_AS is not set
 CONFIG_IOSCHED_DEADLINE=y
 CONFIG_IOSCHED_CFQ=y
 # CONFIG_ATA_OVER_ETH is not set
@@ -308,7 +310,8 @@ CONFIG_BLK_DEV_AMD74XX=y
 CONFIG_BLK_DEV_PIIX=y
 # CONFIG_BLK_DEV_NS87415 is not set
 # CONFIG_BLK_DEV_PDC202XX_OLD is not set
-# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+CONFIG_BLK_DEV_PDC202XX_NEW=y
+# CONFIG_PDC202XX_FORCE is not set
 # CONFIG_BLK_DEV_SVWKS is not set
 # CONFIG_BLK_DEV_SIIMAGE is not set
 # CONFIG_BLK_DEV_SIS5513 is not set
@@ -353,7 +356,7 @@ CONFIG_BLK_DEV_SD=y
 #
 # SCSI low-level drivers
 #
-CONFIG_BLK_DEV_3W_XXXX_RAID=y
+# CONFIG_BLK_DEV_3W_XXXX_RAID is not set
 # CONFIG_SCSI_3W_9XXX is not set
 # CONFIG_SCSI_ACARD is not set
 # CONFIG_SCSI_AACRAID is not set
@@ -384,7 +387,6 @@ CONFIG_SCSI_SATA_VIA=y
 # CONFIG_SCSI_BUSLOGIC is not set
 # CONFIG_SCSI_DMX3191D is not set
 # CONFIG_SCSI_EATA is not set
-# CONFIG_SCSI_EATA_PIO is not set
 # CONFIG_SCSI_FUTURE_DOMAIN is not set
 # CONFIG_SCSI_GDTH is not set
 # CONFIG_SCSI_IPS is not set
@@ -392,7 +394,6 @@ CONFIG_SCSI_SATA_VIA=y
 # CONFIG_SCSI_INIA100 is not set
 # CONFIG_SCSI_SYM53C8XX_2 is not set
 # CONFIG_SCSI_IPR is not set
-# CONFIG_SCSI_QLOGIC_ISP is not set
 # CONFIG_SCSI_QLOGIC_FC is not set
 # CONFIG_SCSI_QLOGIC_1280 is not set
 CONFIG_SCSI_QLA2XXX=y
@@ -401,6 +402,7 @@ CONFIG_SCSI_QLA2XXX=y
 # CONFIG_SCSI_QLA2300 is not set
 # CONFIG_SCSI_QLA2322 is not set
 # CONFIG_SCSI_QLA6312 is not set
+# CONFIG_SCSI_LPFC is not set
 # CONFIG_SCSI_DC395x is not set
 # CONFIG_SCSI_DC390T is not set
 # CONFIG_SCSI_DEBUG is not set
@@ -437,7 +439,6 @@ CONFIG_NET=y
 #
 CONFIG_PACKET=y
 # CONFIG_PACKET_MMAP is not set
-# CONFIG_NETLINK_DEV is not set
 CONFIG_UNIX=y
 # CONFIG_NET_KEY is not set
 CONFIG_INET=y
@@ -502,7 +503,7 @@ CONFIG_NETDEVICES=y
 # CONFIG_DUMMY is not set
 # CONFIG_BONDING is not set
 # CONFIG_EQUALIZER is not set
-# CONFIG_TUN is not set
+CONFIG_TUN=y
 
 #
 # ARCnet devices
@@ -525,8 +526,7 @@ CONFIG_MII=y
 # CONFIG_HP100 is not set
 CONFIG_NET_PCI=y
 # CONFIG_PCNET32 is not set
-CONFIG_AMD8111_ETH=y
-# CONFIG_AMD8111E_NAPI is not set
+# CONFIG_AMD8111_ETH is not set
 # CONFIG_ADAPTEC_STARFIRE is not set
 # CONFIG_B44 is not set
 CONFIG_FORCEDETH=y
@@ -536,7 +536,7 @@ CONFIG_FORCEDETH=y
 # CONFIG_FEALNX is not set
 # CONFIG_NATSEMI is not set
 # CONFIG_NE2K_PCI is not set
-CONFIG_8139CP=m
+CONFIG_8139CP=y
 CONFIG_8139TOO=y
 # CONFIG_8139TOO_PIO is not set
 # CONFIG_8139TOO_TUNE_TWISTER is not set
@@ -671,6 +671,7 @@ CONFIG_SERIAL_8250_NR_UARTS=4
 #
 CONFIG_SERIAL_CORE=y
 CONFIG_SERIAL_CORE_CONSOLE=y
+# CONFIG_SERIAL_JSM is not set
 CONFIG_UNIX98_PTYS=y
 CONFIG_LEGACY_PTYS=y
 CONFIG_LEGACY_PTY_COUNT=256
@@ -696,6 +697,7 @@ CONFIG_RTC=y
 #
 CONFIG_AGP=y
 CONFIG_AGP_AMD64=y
+CONFIG_AGP_INTEL=y
 # CONFIG_DRM is not set
 # CONFIG_MWAVE is not set
 CONFIG_RAW_DRIVER=y
@@ -703,7 +705,7 @@ CONFIG_HPET=y
 # CONFIG_HPET_RTC_IRQ is not set
 CONFIG_HPET_MMAP=y
 CONFIG_MAX_RAW_DEVS=256
-CONFIG_HANGCHECK_TIMER=y
+# CONFIG_HANGCHECK_TIMER is not set
 
 #
 # TPM devices
@@ -786,6 +788,8 @@ CONFIG_SOUND_ICH=y
 #
 # USB support
 #
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
 CONFIG_USB=y
 # CONFIG_USB_DEBUG is not set
 
@@ -797,8 +801,6 @@ CONFIG_USB_DEVICEFS=y
 # CONFIG_USB_DYNAMIC_MINORS is not set
 # CONFIG_USB_SUSPEND is not set
 # CONFIG_USB_OTG is not set
-CONFIG_USB_ARCH_HAS_HCD=y
-CONFIG_USB_ARCH_HAS_OHCI=y
 
 #
 # USB Host Controller Drivers
@@ -826,7 +828,6 @@ CONFIG_USB_PRINTER=y
 #
 CONFIG_USB_STORAGE=y
 # CONFIG_USB_STORAGE_DEBUG is not set
-# CONFIG_USB_STORAGE_RW_DETECT is not set
 # CONFIG_USB_STORAGE_DATAFAB is not set
 # CONFIG_USB_STORAGE_FREECOM is not set
 # CONFIG_USB_STORAGE_ISD200 is not set
@@ -965,7 +966,7 @@ CONFIG_AUTOFS_FS=y
 # CD-ROM/DVD Filesystems
 #
 CONFIG_ISO9660_FS=y
-# CONFIG_JOLIET is not set
+CONFIG_JOLIET=y
 # CONFIG_ZISOFS is not set
 # CONFIG_UDF_FS is not set
 
@@ -1092,9 +1093,10 @@ CONFIG_OPROFILE=y
 #
 # Kernel hacking
 #
+# CONFIG_PRINTK_TIME is not set
 CONFIG_DEBUG_KERNEL=y
 CONFIG_MAGIC_SYSRQ=y
-# CONFIG_PRINTK_TIME is not set
+CONFIG_LOG_BUF_SHIFT=18
 # CONFIG_SCHEDSTATS is not set
 # CONFIG_DEBUG_SLAB is not set
 # CONFIG_DEBUG_SPINLOCK is not set
diff --git a/arch/x86_64/kernel/Makefile b/arch/x86_64/kernel/Makefile
index 0a3318e08ab..5ca4a4598fd 100644
--- a/arch/x86_64/kernel/Makefile
+++ b/arch/x86_64/kernel/Makefile
@@ -28,6 +28,7 @@ obj-$(CONFIG_GART_IOMMU)	+= pci-gart.o aperture.o
 obj-$(CONFIG_DUMMY_IOMMU)	+= pci-nommu.o pci-dma.o
 obj-$(CONFIG_SWIOTLB)		+= swiotlb.o
 obj-$(CONFIG_KPROBES)		+= kprobes.o
+obj-$(CONFIG_X86_PM_TIMER)	+= pmtimer.o
 
 obj-$(CONFIG_MODULES)		+= module.o
 
diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c
index 7e13545748e..f8e6cc4fecd 100644
--- a/arch/x86_64/kernel/apic.c
+++ b/arch/x86_64/kernel/apic.c
@@ -33,6 +33,7 @@
 #include <asm/mpspec.h>
 #include <asm/pgalloc.h>
 #include <asm/mach_apic.h>
+#include <asm/nmi.h>
 
 int apic_verbosity;
 
@@ -925,7 +926,7 @@ __init int oem_force_hpet_timer(void)
 	unsigned id;
 	DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS);
 
-	bitmap_empty(clustermap, NUM_APIC_CLUSTERS);
+	bitmap_zero(clustermap, NUM_APIC_CLUSTERS);
 
 	for (i = 0; i < NR_CPUS; i++) {
 		id = bios_cpu_apicid[i];
@@ -1056,7 +1057,7 @@ int __init APIC_init_uniprocessor (void)
 		nr_ioapics = 0;
 #endif
 	setup_boot_APIC_clock();
-
+	check_nmi_watchdog();
 	return 0;
 }
 
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index 1086b5fcac2..28817490fdc 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -220,13 +220,18 @@ sysret_careful:
 	jmp sysret_check
 
 	/* Handle a signal */ 
-	/* edx:	work flags (arg3) */
 sysret_signal:
 	sti
+	testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
+	jz    1f
+
+	/* Really a signal */
+	/* edx:	work flags (arg3) */
 	leaq do_notify_resume(%rip),%rax
 	leaq -ARGOFFSET(%rsp),%rdi # &pt_regs -> arg1
 	xorl %esi,%esi # oldset -> arg2
 	call ptregscall_common
+1:	movl $_TIF_NEED_RESCHED,%edi
 	jmp sysret_check
 	
 	/* Do syscall tracing */
@@ -484,6 +489,8 @@ retint_careful:
 	jmp retint_check
 	
 retint_signal:
+	testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
+	jz    retint_swapgs
 	sti
 	SAVE_REST
 	movq $-1,ORIG_RAX(%rsp) 			
@@ -492,8 +499,8 @@ retint_signal:
 	call do_notify_resume
 	RESTORE_REST
 	cli
+	movl $_TIF_NEED_RESCHED,%edi
 	GET_THREAD_INFO(%rcx)
-	movl $_TIF_WORK_MASK,%edi
 	jmp retint_check
 
 #ifdef CONFIG_PREEMPT
diff --git a/arch/x86_64/kernel/io_apic.c b/arch/x86_64/kernel/io_apic.c
index 60be58617eb..80e9b498c44 100644
--- a/arch/x86_64/kernel/io_apic.c
+++ b/arch/x86_64/kernel/io_apic.c
@@ -42,6 +42,8 @@
 
 int sis_apic_bug; /* not actually supported, dummy for compile */
 
+static int no_timer_check;
+
 static DEFINE_SPINLOCK(ioapic_lock);
 
 /*
@@ -1601,7 +1603,7 @@ static inline void check_timer(void)
 		 * Ok, does IRQ0 through the IOAPIC work?
 		 */
 		unmask_IO_APIC_irq(0);
-		if (timer_irq_works()) {
+		if (!no_timer_check && timer_irq_works()) {
 			nmi_watchdog_default();
 			if (nmi_watchdog == NMI_IO_APIC) {
 				disable_8259A_irq(0);
@@ -1671,6 +1673,13 @@ static inline void check_timer(void)
 	panic("IO-APIC + timer doesn't work! Try using the 'noapic' kernel parameter\n");
 }
 
+static int __init notimercheck(char *s)
+{
+	no_timer_check = 1;
+	return 1;
+}
+__setup("no_timer_check", notimercheck);
+
 /*
  *
  * IRQ's that are handled by the PIC in the MPS IOAPIC case.
@@ -1804,76 +1813,6 @@ device_initcall(ioapic_init_sysfs);
 
 #define IO_APIC_MAX_ID		0xFE
 
-int __init io_apic_get_unique_id (int ioapic, int apic_id)
-{
-	union IO_APIC_reg_00 reg_00;
-	static physid_mask_t apic_id_map;
-	unsigned long flags;
-	int i = 0;
-
-	/*
-	 * The P4 platform supports up to 256 APIC IDs on two separate APIC 
-	 * buses (one for LAPICs, one for IOAPICs), where predecessors only 
-	 * supports up to 16 on one shared APIC bus.
-	 * 
-	 * TBD: Expand LAPIC/IOAPIC support on P4-class systems to take full
-	 *      advantage of new APIC bus architecture.
-	 */
-
-	if (physids_empty(apic_id_map))
-		apic_id_map = phys_cpu_present_map;
-
-	spin_lock_irqsave(&ioapic_lock, flags);
-	reg_00.raw = io_apic_read(ioapic, 0);
-	spin_unlock_irqrestore(&ioapic_lock, flags);
-
-	if (apic_id >= IO_APIC_MAX_ID) {
-		apic_printk(APIC_QUIET, KERN_WARNING "IOAPIC[%d]: Invalid apic_id %d, trying "
-			"%d\n", ioapic, apic_id, reg_00.bits.ID);
-		apic_id = reg_00.bits.ID;
-	}
-
-	/*
-	 * Every APIC in a system must have a unique ID or we get lots of nice 
-	 * 'stuck on smp_invalidate_needed IPI wait' messages.
-	 */
-	if (physid_isset(apic_id, apic_id_map)) {
-
-		for (i = 0; i < IO_APIC_MAX_ID; i++) {
-			if (!physid_isset(i, apic_id_map))
-				break;
-		}
-
-		if (i == IO_APIC_MAX_ID)
-			panic("Max apic_id exceeded!\n");
-
-		apic_printk(APIC_VERBOSE, KERN_WARNING "IOAPIC[%d]: apic_id %d already used, "
-			"trying %d\n", ioapic, apic_id, i);
-
-		apic_id = i;
-	} 
-
-	physid_set(apic_id, apic_id_map);
-
-	if (reg_00.bits.ID != apic_id) {
-		reg_00.bits.ID = apic_id;
-
-		spin_lock_irqsave(&ioapic_lock, flags);
-		io_apic_write(ioapic, 0, reg_00.raw);
-		reg_00.raw = io_apic_read(ioapic, 0);
-		spin_unlock_irqrestore(&ioapic_lock, flags);
-
-		/* Sanity check */
-		if (reg_00.bits.ID != apic_id)
-			panic("IOAPIC[%d]: Unable change apic_id!\n", ioapic);
-	}
-
-	apic_printk(APIC_VERBOSE,KERN_INFO "IOAPIC[%d]: Assigned apic_id %d\n", ioapic, apic_id);
-
-	return apic_id;
-}
-
-
 int __init io_apic_get_version (int ioapic)
 {
 	union IO_APIC_reg_01	reg_01;
diff --git a/arch/x86_64/kernel/mpparse.c b/arch/x86_64/kernel/mpparse.c
index 7ec031c6ca1..f86d9db94bf 100644
--- a/arch/x86_64/kernel/mpparse.c
+++ b/arch/x86_64/kernel/mpparse.c
@@ -107,6 +107,7 @@ static int __init mpf_checksum(unsigned char *mp, int len)
 static void __init MP_processor_info (struct mpc_config_processor *m)
 {
 	int ver;
+	static int found_bsp=0;
 
 	if (!(m->mpc_cpuflag & CPU_ENABLED))
 		return;
@@ -126,11 +127,6 @@ static void __init MP_processor_info (struct mpc_config_processor *m)
 			" Processor ignored.\n", NR_CPUS);
 		return;
 	}
-	if (num_processors >= maxcpus) {
-		printk(KERN_WARNING "WARNING: maxcpus limit of %i reached."
-			" Processor ignored.\n", maxcpus);
-		return;
-	}
 
 	num_processors++;
 
@@ -150,7 +146,19 @@ static void __init MP_processor_info (struct mpc_config_processor *m)
 		ver = 0x10;
 	}
 	apic_version[m->mpc_apicid] = ver;
-	bios_cpu_apicid[num_processors - 1] = m->mpc_apicid;
+ 	if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) {
+ 		/*
+ 		 * bios_cpu_apicid is required to have processors listed
+ 		 * in same order as logical cpu numbers. Hence the first
+ 		 * entry is BSP, and so on.
+ 		 */
+ 		bios_cpu_apicid[0] = m->mpc_apicid;
+ 		x86_cpu_to_apicid[0] = m->mpc_apicid;
+ 		found_bsp = 1;
+ 	} else {
+ 		bios_cpu_apicid[num_processors - found_bsp] = m->mpc_apicid;
+ 		x86_cpu_to_apicid[num_processors - found_bsp] = m->mpc_apicid;
+ 	}
 }
 
 static void __init MP_bus_info (struct mpc_config_bus *m)
@@ -759,7 +767,7 @@ void __init mp_register_ioapic (
 	mp_ioapics[idx].mpc_apicaddr = address;
 
 	set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address);
-	mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx, id);
+	mp_ioapics[idx].mpc_apicid = id;
 	mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx);
 	
 	/* 
diff --git a/arch/x86_64/kernel/nmi.c b/arch/x86_64/kernel/nmi.c
index 61de0b34a01..31c0f2e6ac9 100644
--- a/arch/x86_64/kernel/nmi.c
+++ b/arch/x86_64/kernel/nmi.c
@@ -33,6 +33,7 @@
 #include <asm/msr.h>
 #include <asm/proto.h>
 #include <asm/kdebug.h>
+#include <asm/local.h>
 
 /*
  * lapic_nmi_owner tracks the ownership of the lapic NMI hardware:
@@ -59,7 +60,8 @@ int panic_on_timeout;
 
 unsigned int nmi_watchdog = NMI_DEFAULT;
 static unsigned int nmi_hz = HZ;
-unsigned int nmi_perfctr_msr;	/* the MSR to reset in NMI handler */
+static unsigned int nmi_perfctr_msr;	/* the MSR to reset in NMI handler */
+static unsigned int nmi_p4_cccr_val;
 
 /* Note that these events don't tick when the CPU idles. This means
    the frequency varies with CPU load. */
@@ -71,61 +73,87 @@ unsigned int nmi_perfctr_msr;	/* the MSR to reset in NMI handler */
 #define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING	0x76
 #define K7_NMI_EVENT		K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING
 
-#define P6_EVNTSEL0_ENABLE	(1 << 22)
-#define P6_EVNTSEL_INT		(1 << 20)
-#define P6_EVNTSEL_OS		(1 << 17)
-#define P6_EVNTSEL_USR		(1 << 16)
-#define P6_EVENT_CPU_CLOCKS_NOT_HALTED	0x79
-#define P6_NMI_EVENT		P6_EVENT_CPU_CLOCKS_NOT_HALTED
+#define MSR_P4_MISC_ENABLE	0x1A0
+#define MSR_P4_MISC_ENABLE_PERF_AVAIL	(1<<7)
+#define MSR_P4_MISC_ENABLE_PEBS_UNAVAIL	(1<<12)
+#define MSR_P4_PERFCTR0		0x300
+#define MSR_P4_CCCR0		0x360
+#define P4_ESCR_EVENT_SELECT(N)	((N)<<25)
+#define P4_ESCR_OS		(1<<3)
+#define P4_ESCR_USR		(1<<2)
+#define P4_CCCR_OVF_PMI0	(1<<26)
+#define P4_CCCR_OVF_PMI1	(1<<27)
+#define P4_CCCR_THRESHOLD(N)	((N)<<20)
+#define P4_CCCR_COMPLEMENT	(1<<19)
+#define P4_CCCR_COMPARE		(1<<18)
+#define P4_CCCR_REQUIRED	(3<<16)
+#define P4_CCCR_ESCR_SELECT(N)	((N)<<13)
+#define P4_CCCR_ENABLE		(1<<12)
+/* Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter
+   CRU_ESCR0 (with any non-null event selector) through a complemented
+   max threshold. [IA32-Vol3, Section 14.9.9] */
+#define MSR_P4_IQ_COUNTER0	0x30C
+#define P4_NMI_CRU_ESCR0	(P4_ESCR_EVENT_SELECT(0x3F)|P4_ESCR_OS|P4_ESCR_USR)
+#define P4_NMI_IQ_CCCR0	\
+	(P4_CCCR_OVF_PMI0|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT|	\
+	 P4_CCCR_COMPARE|P4_CCCR_REQUIRED|P4_CCCR_ESCR_SELECT(4)|P4_CCCR_ENABLE)
+
+static __init inline int nmi_known_cpu(void)
+{
+	switch (boot_cpu_data.x86_vendor) {
+	case X86_VENDOR_AMD:
+		return boot_cpu_data.x86 == 15;
+	case X86_VENDOR_INTEL:
+		return boot_cpu_data.x86 == 15;
+	}
+	return 0;
+}
 
 /* Run after command line and cpu_init init, but before all other checks */
 void __init nmi_watchdog_default(void)
 {
 	if (nmi_watchdog != NMI_DEFAULT)
 		return;
-
-	/* For some reason the IO APIC watchdog doesn't work on the AMD
-	   8111 chipset. For now switch to local APIC mode using
-	   perfctr0 there.  On Intel CPUs we don't have code to handle
-	   the perfctr and the IO-APIC seems to work, so use that.  */
-
-	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
-		nmi_watchdog = NMI_LOCAL_APIC; 
-		printk(KERN_INFO 
-              "Using local APIC NMI watchdog using perfctr0\n");
-	} else {
-		printk(KERN_INFO "Using IO APIC NMI watchdog\n");
+	if (nmi_known_cpu())
+		nmi_watchdog = NMI_LOCAL_APIC;
+	else
 		nmi_watchdog = NMI_IO_APIC;
-	}
 }
 
-/* Why is there no CPUID flag for this? */
-static __init int cpu_has_lapic(void)
+#ifdef CONFIG_SMP
+/* The performance counters used by NMI_LOCAL_APIC don't trigger when
+ * the CPU is idle. To make sure the NMI watchdog really ticks on all
+ * CPUs during the test make them busy.
+ */
+static __init void nmi_cpu_busy(void *data)
 {
-	switch (boot_cpu_data.x86_vendor) { 
-	case X86_VENDOR_INTEL:
-	case X86_VENDOR_AMD: 
-		return boot_cpu_data.x86 >= 6; 
-	/* .... add more cpus here or find a different way to figure this out. */	
-	default:
-		return 0;
-	} 	
+	volatile int *endflag = data;
+	local_irq_enable();
+	/* Intentionally don't use cpu_relax here. This is
+	   to make sure that the performance counter really ticks,
+	   even if there is a simulator or similar that catches the
+	   pause instruction. On a real HT machine this is fine because
+	   all other CPUs are busy with "useless" delay loops and don't
+	   care if they get somewhat less cycles. */
+	while (*endflag == 0)
+		barrier();
 }
+#endif
 
-static int __init check_nmi_watchdog (void)
+int __init check_nmi_watchdog (void)
 {
-	int counts[NR_CPUS];
+	volatile int endflag = 0;
+	int *counts;
 	int cpu;
 
-	if (nmi_watchdog == NMI_NONE)
-		return 0;
+	counts = kmalloc(NR_CPUS * sizeof(int), GFP_KERNEL);
+	if (!counts)
+		return -1;
 
-	if (nmi_watchdog == NMI_LOCAL_APIC && !cpu_has_lapic())  {
-		nmi_watchdog = NMI_NONE;
-		return -1; 
-	}	
+	printk(KERN_INFO "testing NMI watchdog ... ");
 
-	printk(KERN_INFO "Testing NMI watchdog ... ");
+	if (nmi_watchdog == NMI_LOCAL_APIC)
+		smp_call_function(nmi_cpu_busy, (void *)&endflag, 0, 0);
 
 	for (cpu = 0; cpu < NR_CPUS; cpu++)
 		counts[cpu] = cpu_pda[cpu].__nmi_count; 
@@ -133,15 +161,22 @@ static int __init check_nmi_watchdog (void)
 	mdelay((10*1000)/nmi_hz); // wait 10 ticks
 
 	for (cpu = 0; cpu < NR_CPUS; cpu++) {
+		if (!cpu_online(cpu))
+			continue;
 		if (cpu_pda[cpu].__nmi_count - counts[cpu] <= 5) {
-			printk("CPU#%d: NMI appears to be stuck (%d)!\n", 
+			endflag = 1;
+			printk("CPU#%d: NMI appears to be stuck (%d->%d)!\n",
 			       cpu,
+			       counts[cpu],
 			       cpu_pda[cpu].__nmi_count);
 			nmi_active = 0;
 			lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG;
+			nmi_perfctr_msr = 0;
+			kfree(counts);
 			return -1;
 		}
 	}
+	endflag = 1;
 	printk("OK.\n");
 
 	/* now that we know it works we can reduce NMI frequency to
@@ -149,10 +184,9 @@ static int __init check_nmi_watchdog (void)
 	if (nmi_watchdog == NMI_LOCAL_APIC)
 		nmi_hz = 1;
 
+	kfree(counts);
 	return 0;
 }
-/* Have this called later during boot so counters are updating */
-late_initcall(check_nmi_watchdog);
 
 int __init setup_nmi_watchdog(char *str)
 {
@@ -170,7 +204,7 @@ int __init setup_nmi_watchdog(char *str)
 
 	if (nmi >= NMI_INVALID)
 		return 0;
-		nmi_watchdog = nmi;
+	nmi_watchdog = nmi;
 	return 1;
 }
 
@@ -185,7 +219,10 @@ static void disable_lapic_nmi_watchdog(void)
 		wrmsr(MSR_K7_EVNTSEL0, 0, 0);
 		break;
 	case X86_VENDOR_INTEL:
-		wrmsr(MSR_IA32_EVNTSEL0, 0, 0);
+		if (boot_cpu_data.x86 == 15) {
+			wrmsr(MSR_P4_IQ_CCCR0, 0, 0);
+			wrmsr(MSR_P4_CRU_ESCR0, 0, 0);
+		}
 		break;
 	}
 	nmi_active = -1;
@@ -253,7 +290,7 @@ void enable_timer_nmi_watchdog(void)
 
 static int nmi_pm_active; /* nmi_active before suspend */
 
-static int lapic_nmi_suspend(struct sys_device *dev, pm_message_t state)
+static int lapic_nmi_suspend(struct sys_device *dev, u32 state)
 {
 	nmi_pm_active = nmi_active;
 	disable_lapic_nmi_watchdog();
@@ -300,22 +337,27 @@ late_initcall(init_lapic_nmi_sysfs);
  * Original code written by Keith Owens.
  */
 
+static void clear_msr_range(unsigned int base, unsigned int n)
+{
+	unsigned int i;
+
+	for(i = 0; i < n; ++i)
+		wrmsr(base+i, 0, 0);
+}
+
 static void setup_k7_watchdog(void)
 {
 	int i;
 	unsigned int evntsel;
 
-	/* No check, so can start with slow frequency */
-	nmi_hz = 1; 
-
-	/* XXX should check these in EFER */
-
 	nmi_perfctr_msr = MSR_K7_PERFCTR0;
 
 	for(i = 0; i < 4; ++i) {
 		/* Simulator may not support it */
-		if (checking_wrmsrl(MSR_K7_EVNTSEL0+i, 0UL))
+		if (checking_wrmsrl(MSR_K7_EVNTSEL0+i, 0UL)) {
+			nmi_perfctr_msr = 0;
 			return;
+		}
 		wrmsrl(MSR_K7_PERFCTR0+i, 0UL);
 	}
 
@@ -325,12 +367,54 @@ static void setup_k7_watchdog(void)
 		| K7_NMI_EVENT;
 
 	wrmsr(MSR_K7_EVNTSEL0, evntsel, 0);
-	wrmsrl(MSR_K7_PERFCTR0, -((u64)cpu_khz*1000) / nmi_hz);
+	wrmsr(MSR_K7_PERFCTR0, -(cpu_khz/nmi_hz*1000), -1);
 	apic_write(APIC_LVTPC, APIC_DM_NMI);
 	evntsel |= K7_EVNTSEL_ENABLE;
 	wrmsr(MSR_K7_EVNTSEL0, evntsel, 0);
 }
 
+
+static int setup_p4_watchdog(void)
+{
+	unsigned int misc_enable, dummy;
+
+	rdmsr(MSR_P4_MISC_ENABLE, misc_enable, dummy);
+	if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL))
+		return 0;
+
+	nmi_perfctr_msr = MSR_P4_IQ_COUNTER0;
+	nmi_p4_cccr_val = P4_NMI_IQ_CCCR0;
+#ifdef CONFIG_SMP
+	if (smp_num_siblings == 2)
+		nmi_p4_cccr_val |= P4_CCCR_OVF_PMI1;
+#endif
+
+	if (!(misc_enable & MSR_P4_MISC_ENABLE_PEBS_UNAVAIL))
+		clear_msr_range(0x3F1, 2);
+	/* MSR 0x3F0 seems to have a default value of 0xFC00, but current
+	   docs doesn't fully define it, so leave it alone for now. */
+	if (boot_cpu_data.x86_model >= 0x3) {
+		/* MSR_P4_IQ_ESCR0/1 (0x3ba/0x3bb) removed */
+		clear_msr_range(0x3A0, 26);
+		clear_msr_range(0x3BC, 3);
+	} else {
+		clear_msr_range(0x3A0, 31);
+	}
+	clear_msr_range(0x3C0, 6);
+	clear_msr_range(0x3C8, 6);
+	clear_msr_range(0x3E0, 2);
+	clear_msr_range(MSR_P4_CCCR0, 18);
+	clear_msr_range(MSR_P4_PERFCTR0, 18);
+
+	wrmsr(MSR_P4_CRU_ESCR0, P4_NMI_CRU_ESCR0, 0);
+	wrmsr(MSR_P4_IQ_CCCR0, P4_NMI_IQ_CCCR0 & ~P4_CCCR_ENABLE, 0);
+	Dprintk("setting P4_IQ_COUNTER0 to 0x%08lx\n", -(cpu_khz/nmi_hz*1000));
+	wrmsr(MSR_P4_IQ_COUNTER0, -(cpu_khz/nmi_hz*1000), -1);
+	apic_write(APIC_LVTPC, APIC_DM_NMI);
+	wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0);
+	return 1;
+}
+
 void setup_apic_nmi_watchdog(void)
 {
 	switch (boot_cpu_data.x86_vendor) {
@@ -341,6 +425,13 @@ void setup_apic_nmi_watchdog(void)
 			return;
 		setup_k7_watchdog();
 		break;
+	case X86_VENDOR_INTEL:
+		if (boot_cpu_data.x86 != 15)
+			return;
+		if (!setup_p4_watchdog())
+			return;
+		break;
+
 	default:
 		return;
 	}
@@ -355,56 +446,67 @@ void setup_apic_nmi_watchdog(void)
  *
  * as these watchdog NMI IRQs are generated on every CPU, we only
  * have to check the current processor.
- *
- * since NMIs don't listen to _any_ locks, we have to be extremely
- * careful not to rely on unsafe variables. The printk might lock
- * up though, so we have to break up any console locks first ...
- * [when there will be more tty-related locks, break them up
- *  here too!]
  */
 
-static unsigned int
-	last_irq_sums [NR_CPUS],
-	alert_counter [NR_CPUS];
+static DEFINE_PER_CPU(unsigned, last_irq_sum);
+static DEFINE_PER_CPU(local_t, alert_counter);
+static DEFINE_PER_CPU(int, nmi_touch);
 
 void touch_nmi_watchdog (void)
 {
 	int i;
 
 	/*
-	 * Just reset the alert counters, (other CPUs might be
-	 * spinning on locks we hold):
+ 	 * Tell other CPUs to reset their alert counters. We cannot
+	 * do it ourselves because the alert count increase is not
+	 * atomic.
 	 */
 	for (i = 0; i < NR_CPUS; i++)
-		alert_counter[i] = 0;
+		per_cpu(nmi_touch, i) = 1;
 }
 
 void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason)
 {
-	int sum, cpu;
+	int sum;
+	int touched = 0;
 
-	cpu = safe_smp_processor_id();
 	sum = read_pda(apic_timer_irqs);
-	if (last_irq_sums[cpu] == sum) {
+	if (__get_cpu_var(nmi_touch)) {
+		__get_cpu_var(nmi_touch) = 0;
+		touched = 1;
+	}
+	if (!touched && __get_cpu_var(last_irq_sum) == sum) {
 		/*
 		 * Ayiee, looks like this CPU is stuck ...
 		 * wait a few IRQs (5 seconds) before doing the oops ...
 		 */
-		alert_counter[cpu]++;
-		if (alert_counter[cpu] == 5*nmi_hz) {
+		local_inc(&__get_cpu_var(alert_counter));
+		if (local_read(&__get_cpu_var(alert_counter)) == 5*nmi_hz) {
 			if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT)
 							== NOTIFY_STOP) {
-				alert_counter[cpu] = 0; 
+				local_set(&__get_cpu_var(alert_counter), 0);
 				return;
 			} 
 			die_nmi("NMI Watchdog detected LOCKUP on CPU%d", regs);
 		}
 	} else {
-		last_irq_sums[cpu] = sum;
-		alert_counter[cpu] = 0;
+		__get_cpu_var(last_irq_sum) = sum;
+		local_set(&__get_cpu_var(alert_counter), 0);
 	}
-	if (nmi_perfctr_msr)
+	if (nmi_perfctr_msr) {
+ 		if (nmi_perfctr_msr == MSR_P4_IQ_COUNTER0) {
+ 			/*
+ 			 * P4 quirks:
+ 			 * - An overflown perfctr will assert its interrupt
+ 			 *   until the OVF flag in its CCCR is cleared.
+ 			 * - LVTPC is masked on interrupt and must be
+ 			 *   unmasked by the LVTPC handler.
+ 			 */
+ 			wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0);
+ 			apic_write(APIC_LVTPC, APIC_DM_NMI);
+ 		}
 		wrmsr(nmi_perfctr_msr, -(cpu_khz/nmi_hz*1000), -1);
+	}
 }
 
 static int dummy_nmi_callback(struct pt_regs * regs, int cpu)
diff --git a/arch/x86_64/kernel/pmtimer.c b/arch/x86_64/kernel/pmtimer.c
new file mode 100644
index 00000000000..feb5f108dd2
--- /dev/null
+++ b/arch/x86_64/kernel/pmtimer.c
@@ -0,0 +1,101 @@
+/* Ported over from i386 by AK, original copyright was:
+ *
+ * (C) Dominik Brodowski <linux@brodo.de> 2003
+ *
+ * Driver to use the Power Management Timer (PMTMR) available in some
+ * southbridges as primary timing source for the Linux kernel.
+ *
+ * Based on parts of linux/drivers/acpi/hardware/hwtimer.c, timer_pit.c,
+ * timer_hpet.c, and on Arjan van de Ven's implementation for 2.4.
+ *
+ * This file is licensed under the GPL v2.
+ *
+ * Dropped all the hardware bug workarounds for now. Hopefully they
+ * are not needed on 64bit chipsets.
+ */
+
+#include <linux/jiffies.h>
+#include <linux/kernel.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/cpumask.h>
+#include <asm/io.h>
+#include <asm/proto.h>
+#include <asm/msr.h>
+#include <asm/vsyscall.h>
+
+/* The I/O port the PMTMR resides at.
+ * The location is detected during setup_arch(),
+ * in arch/i386/kernel/acpi/boot.c */
+u32 pmtmr_ioport;
+
+/* value of the Power timer at last timer interrupt */
+static u32 offset_delay;
+static u32 last_pmtmr_tick;
+
+#define ACPI_PM_MASK 0xFFFFFF /* limit it to 24 bits */
+
+static inline u32 cyc2us(u32 cycles)
+{
+	/* The Power Management Timer ticks at 3.579545 ticks per microsecond.
+	 * 1 / PM_TIMER_FREQUENCY == 0.27936511 =~ 286/1024 [error: 0.024%]
+	 *
+	 * Even with HZ = 100, delta is at maximum 35796 ticks, so it can
+	 * easily be multiplied with 286 (=0x11E) without having to fear
+	 * u32 overflows.
+	 */
+	cycles *= 286;
+	return (cycles >> 10);
+}
+
+int pmtimer_mark_offset(void)
+{
+	static int first_run = 1;
+	unsigned long tsc;
+	u32 lost;
+
+	u32 tick = inl(pmtmr_ioport);
+	u32 delta;
+
+	delta = cyc2us((tick - last_pmtmr_tick) & ACPI_PM_MASK);
+
+	last_pmtmr_tick = tick;
+	monotonic_base += delta * NSEC_PER_USEC;
+
+	delta += offset_delay;
+
+	lost = delta / (USEC_PER_SEC / HZ);
+	offset_delay = delta % (USEC_PER_SEC / HZ);
+
+	rdtscll(tsc);
+	vxtime.last_tsc = tsc - offset_delay * cpu_khz;
+
+	/* don't calculate delay for first run,
+	   or if we've got less then a tick */
+	if (first_run || (lost < 1)) {
+		first_run = 0;
+		offset_delay = 0;
+	}
+
+	return lost - 1;
+}
+
+unsigned int do_gettimeoffset_pm(void)
+{
+	u32 now, offset, delta = 0;
+
+	offset = last_pmtmr_tick;
+	now = inl(pmtmr_ioport);
+	delta = (now - offset) & ACPI_PM_MASK;
+
+	return offset_delay + cyc2us(delta);
+}
+
+
+static int __init nopmtimer_setup(char *s)
+{
+	pmtmr_ioport = 0;
+	return 0;
+}
+
+__setup("nopmtimer", nopmtimer_setup);
diff --git a/arch/x86_64/kernel/ptrace.c b/arch/x86_64/kernel/ptrace.c
index e26e86bb56f..525f6a128a2 100644
--- a/arch/x86_64/kernel/ptrace.c
+++ b/arch/x86_64/kernel/ptrace.c
@@ -257,13 +257,13 @@ static int putreg(struct task_struct *child,
 			value &= 0xffff;
 			return 0;
 		case offsetof(struct user_regs_struct,fs_base):
-			if (!((value >> 48) == 0 || (value >> 48) == 0xffff))
-				return -EIO; 
+			if (value >= TASK_SIZE)
+				return -EIO;
 			child->thread.fs = value;
 			return 0;
 		case offsetof(struct user_regs_struct,gs_base):
-			if (!((value >> 48) == 0 || (value >> 48) == 0xffff))
-				return -EIO; 
+			if (value >= TASK_SIZE)
+				return -EIO;
 			child->thread.gs = value;
 			return 0;
 		case offsetof(struct user_regs_struct, eflags):
@@ -277,6 +277,11 @@ static int putreg(struct task_struct *child,
 				return -EIO;
 			value &= 0xffff;
 			break;
+		case offsetof(struct user_regs_struct, rip):
+			/* Check if the new RIP address is canonical */
+			if (value >= TASK_SIZE)
+				return -EIO;
+			break;
 	}
 	put_stack_long(child, regno - sizeof(struct pt_regs), value);
 	return 0;
@@ -375,7 +380,7 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data
 			break;
 
 		switch (addr) { 
-		case 0 ... sizeof(struct user_regs_struct):
+		case 0 ... sizeof(struct user_regs_struct) - sizeof(long):
 			tmp = getreg(child, addr);
 			break;
 		case offsetof(struct user, u_debugreg[0]):
@@ -420,7 +425,7 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data
 			break;
 
 		switch (addr) { 
-		case 0 ... sizeof(struct user_regs_struct): 
+		case 0 ... sizeof(struct user_regs_struct) - sizeof(long):
 			ret = putreg(child, addr, data);
 			break;
 		/* Disallows to set a breakpoint into the vsyscall */
diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c
index 2129cf9ba6b..99f038ede23 100644
--- a/arch/x86_64/kernel/setup.c
+++ b/arch/x86_64/kernel/setup.c
@@ -719,7 +719,6 @@ static void __init display_cacheinfo(struct cpuinfo_x86 *c)
 	}
 }
 
-#ifdef CONFIG_SMP
 /*
  * On a AMD dual core setup the lower bits of the APIC id distingush the cores.
  * Assumes number of cores is a power of two.
@@ -727,17 +726,26 @@ static void __init display_cacheinfo(struct cpuinfo_x86 *c)
 static void __init amd_detect_cmp(struct cpuinfo_x86 *c)
 {
 #ifdef CONFIG_SMP
-	int cpu = c->x86_apicid;
+	int cpu = smp_processor_id();
 	int node = 0;
+	unsigned bits;
 	if (c->x86_num_cores == 1)
 		return;
-	cpu_core_id[cpu] = cpu >> hweight32(c->x86_num_cores - 1);
+
+	bits = 0;
+	while ((1 << bits) < c->x86_num_cores)
+		bits++;
+
+	/* Low order bits define the core id (index of core in socket) */
+	cpu_core_id[cpu] = phys_proc_id[cpu] & ((1 << bits)-1);
+	/* Convert the APIC ID into the socket ID */
+	phys_proc_id[cpu] >>= bits;
 
 #ifdef CONFIG_NUMA
 	/* When an ACPI SRAT table is available use the mappings from SRAT
  	   instead. */
 	if (acpi_numa <= 0) {
-		node = cpu_core_id[cpu];
+		node = phys_proc_id[cpu];
 		if (!node_online(node))
 			node = first_node(node_online_map);
 		cpu_to_node[cpu] = node;
@@ -745,15 +753,11 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c)
 		node = cpu_to_node[cpu];
 	}
 #endif
+
 	printk(KERN_INFO "CPU %d(%d) -> Node %d -> Core %d\n",
 			cpu, c->x86_num_cores, node, cpu_core_id[cpu]);
 #endif
 }
-#else
-static void __init amd_detect_cmp(struct cpuinfo_x86 *c)
-{
-}
-#endif
 
 static int __init init_amd(struct cpuinfo_x86 *c)
 {
@@ -925,7 +929,6 @@ void __init early_identify_cpu(struct cpuinfo_x86 *c)
 	c->x86_clflush_size = 64;
 	c->x86_cache_alignment = c->x86_clflush_size;
 	c->x86_num_cores = 1;
-	c->x86_apicid = c == &boot_cpu_data ? 0 : c - cpu_data;
 	c->extended_cpuid_level = 0;
 	memset(&c->x86_capability, 0, sizeof c->x86_capability);
 
@@ -954,11 +957,14 @@ void __init early_identify_cpu(struct cpuinfo_x86 *c)
 		} 
 		if (c->x86_capability[0] & (1<<19)) 
 			c->x86_clflush_size = ((misc >> 8) & 0xff) * 8;
-		c->x86_apicid = misc >> 24;
 	} else {
 		/* Have CPUID level 0 only - unheard of */
 		c->x86 = 4;
 	}
+
+#ifdef CONFIG_SMP
+	phys_proc_id[smp_processor_id()] = (cpuid_ebx(1) >> 24) & 0xff;
+#endif
 }
 
 /*
@@ -1088,7 +1094,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
 		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
 
 		/* Other (Linux-defined) */
-		"cxmmx", NULL, "cyrix_arr", "centaur_mcr", "k8c+",
+		"cxmmx", NULL, "cyrix_arr", "centaur_mcr", NULL,
 		"constant_tsc", NULL, NULL,
 		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
 		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
diff --git a/arch/x86_64/kernel/signal.c b/arch/x86_64/kernel/signal.c
index d439ced150c..3fdcdba0fec 100644
--- a/arch/x86_64/kernel/signal.c
+++ b/arch/x86_64/kernel/signal.c
@@ -452,7 +452,9 @@ int do_signal(struct pt_regs *regs, sigset_t *oldset)
 			regs->rip -= 2;
 		}
 		if (regs->rax == (unsigned long)-ERESTART_RESTARTBLOCK) {
-			regs->rax = __NR_restart_syscall;
+			regs->rax = test_thread_flag(TIF_IA32) ?
+					__NR_ia32_restart_syscall :
+					__NR_restart_syscall;
 			regs->rip -= 2;
 		}
 	}
diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c
index 73f7e8b9543..f1ec0f34594 100644
--- a/arch/x86_64/kernel/smpboot.c
+++ b/arch/x86_64/kernel/smpboot.c
@@ -56,6 +56,7 @@
 #include <asm/kdebug.h>
 #include <asm/tlbflush.h>
 #include <asm/proto.h>
+#include <asm/nmi.h>
 
 /* Change for real CPU hotplug. Note other files need to be fixed
    first too. */
@@ -93,6 +94,7 @@ int smp_threads_ready;
 
 cpumask_t cpu_sibling_map[NR_CPUS] __cacheline_aligned;
 cpumask_t cpu_core_map[NR_CPUS] __cacheline_aligned;
+EXPORT_SYMBOL(cpu_core_map);
 
 /*
  * Trampoline 80x86 program as an array.
@@ -125,96 +127,210 @@ static void __cpuinit smp_store_cpu_info(int id)
 
 	*c = boot_cpu_data;
 	identify_cpu(c);
+	print_cpu_info(c);
 }
 
 /*
- * Synchronize TSCs of CPUs
+ * New Funky TSC sync algorithm borrowed from IA64.
+ * Main advantage is that it doesn't reset the TSCs fully and
+ * in general looks more robust and it works better than my earlier
+ * attempts. I believe it was written by David Mosberger. Some minor
+ * adjustments for x86-64 by me -AK
  *
- * This new algorithm is less accurate than the old "zero TSCs"
- * one, but we cannot zero TSCs anymore in the new hotplug CPU
- * model.
+ * Original comment reproduced below.
+ *
+ * Synchronize TSC of the current (slave) CPU with the TSC of the
+ * MASTER CPU (normally the time-keeper CPU).  We use a closed loop to
+ * eliminate the possibility of unaccounted-for errors (such as
+ * getting a machine check in the middle of a calibration step).  The
+ * basic idea is for the slave to ask the master what itc value it has
+ * and to read its own itc before and after the master responds.  Each
+ * iteration gives us three timestamps:
+ *
+ *	slave		master
+ *
+ *	t0 ---\
+ *             ---\
+ *		   --->
+ *			tm
+ *		   /---
+ *	       /---
+ *	t1 <---
+ *
+ *
+ * The goal is to adjust the slave's TSC such that tm falls exactly
+ * half-way between t0 and t1.  If we achieve this, the clocks are
+ * synchronized provided the interconnect between the slave and the
+ * master is symmetric.  Even if the interconnect were asymmetric, we
+ * would still know that the synchronization error is smaller than the
+ * roundtrip latency (t0 - t1).
+ *
+ * When the interconnect is quiet and symmetric, this lets us
+ * synchronize the TSC to within one or two cycles.  However, we can
+ * only *guarantee* that the synchronization is accurate to within a
+ * round-trip time, which is typically in the range of several hundred
+ * cycles (e.g., ~500 cycles).  In practice, this means that the TSCs
+ * are usually almost perfectly synchronized, but we shouldn't assume
+ * that the accuracy is much better than half a micro second or so.
+ *
+ * [there are other errors like the latency of RDTSC and of the
+ * WRMSR. These can also account to hundreds of cycles. So it's
+ * probably worse. It claims 153 cycles error on a dual Opteron,
+ * but I suspect the numbers are actually somewhat worse -AK]
  */
 
-static atomic_t __cpuinitdata tsc_flag;
+#define MASTER	0
+#define SLAVE	(SMP_CACHE_BYTES/8)
+
+/* Intentionally don't use cpu_relax() while TSC synchronization
+   because we don't want to go into funky power save modi or cause
+   hypervisors to schedule us away.  Going to sleep would likely affect
+   latency and low latency is the primary objective here. -AK */
+#define no_cpu_relax() barrier()
+
 static __cpuinitdata DEFINE_SPINLOCK(tsc_sync_lock);
-static unsigned long long __cpuinitdata bp_tsc, ap_tsc;
+static volatile __cpuinitdata unsigned long go[SLAVE + 1];
+static int notscsync __cpuinitdata;
+
+#undef DEBUG_TSC_SYNC
 
-#define NR_LOOPS 5
+#define NUM_ROUNDS	64	/* magic value */
+#define NUM_ITERS	5	/* likewise */
 
-static void __cpuinit sync_tsc_bp_init(int init)
+/* Callback on boot CPU */
+static __cpuinit void sync_master(void *arg)
 {
-	if (init)
-		_raw_spin_lock(&tsc_sync_lock);
-	else
-		_raw_spin_unlock(&tsc_sync_lock);
-	atomic_set(&tsc_flag, 0);
+	unsigned long flags, i;
+
+	if (smp_processor_id() != boot_cpu_id)
+		return;
+
+	go[MASTER] = 0;
+
+	local_irq_save(flags);
+	{
+		for (i = 0; i < NUM_ROUNDS*NUM_ITERS; ++i) {
+			while (!go[MASTER])
+				no_cpu_relax();
+			go[MASTER] = 0;
+			rdtscll(go[SLAVE]);
+		}
+	}
+	local_irq_restore(flags);
 }
 
 /*
- * Synchronize TSC on AP with BP.
+ * Return the number of cycles by which our tsc differs from the tsc
+ * on the master (time-keeper) CPU.  A positive number indicates our
+ * tsc is ahead of the master, negative that it is behind.
  */
-static void __cpuinit __sync_tsc_ap(void)
+static inline long
+get_delta(long *rt, long *master)
 {
-	if (!cpu_has_tsc)
-		return;
-	Dprintk("AP %d syncing TSC\n", smp_processor_id());
+	unsigned long best_t0 = 0, best_t1 = ~0UL, best_tm = 0;
+	unsigned long tcenter, t0, t1, tm;
+	int i;
 
-	while (atomic_read(&tsc_flag) != 0)
-		cpu_relax();
-	atomic_inc(&tsc_flag);
-	mb();
-	_raw_spin_lock(&tsc_sync_lock);
-	wrmsrl(MSR_IA32_TSC, bp_tsc);
-	_raw_spin_unlock(&tsc_sync_lock);
-	rdtscll(ap_tsc);
-	mb();
-	atomic_inc(&tsc_flag);
-	mb();
+	for (i = 0; i < NUM_ITERS; ++i) {
+		rdtscll(t0);
+		go[MASTER] = 1;
+		while (!(tm = go[SLAVE]))
+			no_cpu_relax();
+		go[SLAVE] = 0;
+		rdtscll(t1);
+
+		if (t1 - t0 < best_t1 - best_t0)
+			best_t0 = t0, best_t1 = t1, best_tm = tm;
+	}
+
+	*rt = best_t1 - best_t0;
+	*master = best_tm - best_t0;
+
+	/* average best_t0 and best_t1 without overflow: */
+	tcenter = (best_t0/2 + best_t1/2);
+	if (best_t0 % 2 + best_t1 % 2 == 2)
+		++tcenter;
+	return tcenter - best_tm;
 }
 
-static void __cpuinit sync_tsc_ap(void)
+static __cpuinit void sync_tsc(void)
 {
-	int i;
-	for (i = 0; i < NR_LOOPS; i++)
-		__sync_tsc_ap();
+	int i, done = 0;
+	long delta, adj, adjust_latency = 0;
+	unsigned long flags, rt, master_time_stamp, bound;
+#if DEBUG_TSC_SYNC
+	static struct syncdebug {
+		long rt;	/* roundtrip time */
+		long master;	/* master's timestamp */
+		long diff;	/* difference between midpoint and master's timestamp */
+		long lat;	/* estimate of tsc adjustment latency */
+	} t[NUM_ROUNDS] __cpuinitdata;
+#endif
+
+	go[MASTER] = 1;
+
+	smp_call_function(sync_master, NULL, 1, 0);
+
+	while (go[MASTER])	/* wait for master to be ready */
+		no_cpu_relax();
+
+	spin_lock_irqsave(&tsc_sync_lock, flags);
+	{
+		for (i = 0; i < NUM_ROUNDS; ++i) {
+			delta = get_delta(&rt, &master_time_stamp);
+			if (delta == 0) {
+				done = 1;	/* let's lock on to this... */
+				bound = rt;
+			}
+
+			if (!done) {
+				unsigned long t;
+				if (i > 0) {
+					adjust_latency += -delta;
+					adj = -delta + adjust_latency/4;
+				} else
+					adj = -delta;
+
+				rdtscll(t);
+				wrmsrl(MSR_IA32_TSC, t + adj);
+			}
+#if DEBUG_TSC_SYNC
+			t[i].rt = rt;
+			t[i].master = master_time_stamp;
+			t[i].diff = delta;
+			t[i].lat = adjust_latency/4;
+#endif
+		}
+	}
+	spin_unlock_irqrestore(&tsc_sync_lock, flags);
+
+#if DEBUG_TSC_SYNC
+	for (i = 0; i < NUM_ROUNDS; ++i)
+		printk("rt=%5ld master=%5ld diff=%5ld adjlat=%5ld\n",
+		       t[i].rt, t[i].master, t[i].diff, t[i].lat);
+#endif
+
+	printk(KERN_INFO
+	       "CPU %d: synchronized TSC with CPU %u (last diff %ld cycles, "
+	       "maxerr %lu cycles)\n",
+	       smp_processor_id(), boot_cpu_id, delta, rt);
 }
 
-/*
- * Synchronize TSC from BP to AP.
- */
-static void __cpuinit __sync_tsc_bp(int cpu)
+static void __cpuinit tsc_sync_wait(void)
 {
-	if (!cpu_has_tsc)
+	if (notscsync || !cpu_has_tsc)
 		return;
-
-	/* Wait for AP */
-	while (atomic_read(&tsc_flag) == 0)
-		cpu_relax();
-	/* Save BPs TSC */
-	sync_core();
-	rdtscll(bp_tsc);
-	/* Don't do the sync core here to avoid too much latency. */
-	mb();
-	/* Start the AP */
-	_raw_spin_unlock(&tsc_sync_lock);
-	/* Wait for AP again */
-	while (atomic_read(&tsc_flag) < 2)
-		cpu_relax();
-	rdtscl(bp_tsc);
-	barrier();
+	printk(KERN_INFO "CPU %d: Syncing TSC to CPU %u.\n", smp_processor_id(),
+			boot_cpu_id);
+	sync_tsc();
 }
 
-static void __cpuinit sync_tsc_bp(int cpu)
+static __init int notscsync_setup(char *s)
 {
-	int i;
-	for (i = 0; i < NR_LOOPS - 1; i++) {
-		__sync_tsc_bp(cpu);
-		sync_tsc_bp_init(1);
-	}
-	__sync_tsc_bp(cpu);
-	printk(KERN_INFO "Synced TSC of CPU %d difference %Ld\n",
-	       cpu, ap_tsc - bp_tsc);
+	notscsync = 1;
+	return 0;
 }
+__setup("notscsync", notscsync_setup);
 
 static atomic_t init_deasserted __cpuinitdata;
 
@@ -315,11 +431,6 @@ void __cpuinit start_secondary(void)
 	cpu_init();
 	smp_callin();
 
-	/*
-	 * Synchronize the TSC with the BP
-	 */
-	sync_tsc_ap();
-
 	/* otherwise gcc will move up the smp_processor_id before the cpu_init */
 	barrier();
 
@@ -334,7 +445,6 @@ void __cpuinit start_secondary(void)
 		enable_8259A_irq(0);
 	}
 
-
 	enable_APIC_timer();
 
 	/*
@@ -343,6 +453,11 @@ void __cpuinit start_secondary(void)
 	cpu_set(smp_processor_id(), cpu_online_map);
 	mb();
 
+	/* Wait for TSC sync to not schedule things before.
+	   We still process interrupts, which could see an inconsistent
+	   time in that window unfortunately. */
+	tsc_sync_wait();
+
 	cpu_idle();
 }
 
@@ -531,7 +646,6 @@ static int __cpuinit do_boot_cpu(int cpu, int apicid)
 		printk("failed fork for CPU %d\n", cpu);
 		return PTR_ERR(idle);
 	}
-	x86_cpu_to_apicid[cpu] = apicid;
 
 	cpu_pda[cpu].pcurrent = idle;
 
@@ -600,8 +714,6 @@ static int __cpuinit do_boot_cpu(int cpu, int apicid)
 
 		if (cpu_isset(cpu, cpu_callin_map)) {
 			/* number CPUs logically, starting from 1 (BSP is 0) */
-			Dprintk("OK.\n");
-			print_cpu_info(&cpu_data[cpu]);
 			Dprintk("CPU has booted.\n");
 		} else {
 			boot_error = 1;
@@ -842,7 +954,6 @@ void __cpuinit smp_prepare_cpus(unsigned int max_cpus)
 		      GET_APIC_ID(apic_read(APIC_ID)), boot_cpu_id);
 		/* Or can we switch back to PIC here? */
 	}
-	x86_cpu_to_apicid[0] = boot_cpu_id;
 
 	/*
 	 * Now start the IO-APICs
@@ -889,18 +1000,14 @@ int __cpuinit __cpu_up(unsigned int cpu)
 		printk("__cpu_up: bad cpu %d\n", cpu);
 		return -EINVAL;
 	}
-	sync_tsc_bp_init(1);
 
 	/* Boot it! */
 	err = do_boot_cpu(cpu, apicid);
 	if (err < 0) {
-		sync_tsc_bp_init(0);
 		Dprintk("do_boot_cpu failed %d\n", err);
 		return err;
 	}
 
-	sync_tsc_bp(cpu);
-
 	/* Unleash the CPU! */
 	Dprintk("waiting for cpu %d\n", cpu);
 
@@ -923,4 +1030,6 @@ void __cpuinit smp_cpus_done(unsigned int max_cpus)
 
 	detect_siblings();
 	time_init_gtod();
+
+	check_nmi_watchdog();
 }
diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c
index 0e3b60e1ac5..735b6767c8e 100644
--- a/arch/x86_64/kernel/time.c
+++ b/arch/x86_64/kernel/time.c
@@ -26,6 +26,8 @@
 #include <linux/sysdev.h>
 #include <linux/bcd.h>
 #include <linux/kallsyms.h>
+#include <linux/acpi.h>
+#include <acpi/achware.h>	/* for PM timer frequency */
 #include <asm/8253pit.h>
 #include <asm/pgtable.h>
 #include <asm/vsyscall.h>
@@ -396,6 +398,10 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
 			(offset - vxtime.last)*(NSEC_PER_SEC/HZ) / hpet_tick;
 
 		vxtime.last = offset;
+#ifdef CONFIG_X86_PM_TIMER
+	} else if (vxtime.mode == VXTIME_PMTMR) {
+		lost = pmtimer_mark_offset();
+#endif
 	} else {
 		offset = (((tsc - vxtime.last_tsc) *
 			   vxtime.tsc_quot) >> 32) - (USEC_PER_SEC / HZ);
@@ -898,6 +904,13 @@ void __init time_init(void)
 			hpet_period;
 		cpu_khz = hpet_calibrate_tsc();
 		timename = "HPET";
+#ifdef CONFIG_X86_PM_TIMER
+	} else if (pmtmr_ioport) {
+		vxtime_hz = PM_TIMER_FREQUENCY;
+		timename = "PM";
+		pit_init();
+		cpu_khz = pit_calibrate_tsc();
+#endif
 	} else {
 		pit_init();
 		cpu_khz = pit_calibrate_tsc();
@@ -923,35 +936,50 @@ void __init time_init(void)
 }
 
 /*
+ * Make an educated guess if the TSC is trustworthy and synchronized
+ * over all CPUs.
+ */
+static __init int unsynchronized_tsc(void)
+{
+#ifdef CONFIG_SMP
+	if (oem_force_hpet_timer())
+		return 1;
+ 	/* Intel systems are normally all synchronized. Exceptions
+ 	   are handled in the OEM check above. */
+ 	if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL)
+ 		return 0;
+ 	/* All in a single socket - should be synchronized */
+ 	if (cpus_weight(cpu_core_map[0]) == num_online_cpus())
+ 		return 0;
+#endif
+ 	/* Assume multi socket systems are not synchronized */
+ 	return num_online_cpus() > 1;
+}
+
+/*
  * Decide after all CPUs are booted what mode gettimeofday should use.
  */
 void __init time_init_gtod(void)
 {
 	char *timetype;
 
-	/*
-	 * AMD systems with more than one CPU don't have fully synchronized
-	 * TSCs. Always use HPET gettimeofday for these, although it is slower.
-	 * Intel SMP systems usually have synchronized TSCs, so use always
-	 * the TSC.
-	 *
-	 * Exceptions:
-	 * IBM Summit2 checked by oem_force_hpet_timer().
- 	 * AMD dual core may also not need HPET. Check me.
-	 *
-	 * Can be turned off with "notsc".
-	 */
-	if (num_online_cpus() > 1 &&
-	    boot_cpu_data.x86_vendor == X86_VENDOR_AMD)
-		notsc = 1;
-	/* Some systems will want to disable TSC and use HPET. */
-	if (oem_force_hpet_timer())
+	if (unsynchronized_tsc())
 		notsc = 1;
 	if (vxtime.hpet_address && notsc) {
 		timetype = "HPET";
 		vxtime.last = hpet_readl(HPET_T0_CMP) - hpet_tick;
 		vxtime.mode = VXTIME_HPET;
 		do_gettimeoffset = do_gettimeoffset_hpet;
+#ifdef CONFIG_X86_PM_TIMER
+	/* Using PM for gettimeofday is quite slow, but we have no other
+	   choice because the TSC is too unreliable on some systems. */
+	} else if (pmtmr_ioport && !vxtime.hpet_address && notsc) {
+		timetype = "PM";
+		do_gettimeoffset = do_gettimeoffset_pm;
+		vxtime.mode = VXTIME_PMTMR;
+		sysctl_vsyscall = 0;
+		printk(KERN_INFO "Disabling vsyscall due to use of PM timer\n");
+#endif
 	} else {
 		timetype = vxtime.hpet_address ? "HPET/TSC" : "PIT/TSC";
 		vxtime.mode = VXTIME_TSC;
diff --git a/arch/x86_64/kernel/traps.c b/arch/x86_64/kernel/traps.c
index 65a37f52c56..3dfec8fdabc 100644
--- a/arch/x86_64/kernel/traps.c
+++ b/arch/x86_64/kernel/traps.c
@@ -332,10 +332,12 @@ void handle_BUG(struct pt_regs *regs)
 	printk(KERN_ALERT "Kernel BUG at %.50s:%d\n", f.filename, f.line);
 } 
 
+#ifdef CONFIG_BUG
 void out_of_line_bug(void)
 { 
 	BUG(); 
 } 
+#endif
 
 static DEFINE_SPINLOCK(die_lock);
 static int die_owner = -1;
diff --git a/arch/x86_64/kernel/vsyscall.c b/arch/x86_64/kernel/vsyscall.c
index b4b8dc59663..2e573442594 100644
--- a/arch/x86_64/kernel/vsyscall.c
+++ b/arch/x86_64/kernel/vsyscall.c
@@ -65,7 +65,7 @@ static force_inline void do_vgettimeofday(struct timeval * tv)
 		usec = (__xtime.tv_nsec / 1000) +
 			(__jiffies - __wall_jiffies) * (1000000 / HZ);
 
-		if (__vxtime.mode == VXTIME_TSC) {
+		if (__vxtime.mode != VXTIME_HPET) {
 			sync_core();
 			rdtscll(t);
 			if (t < __vxtime.last_tsc)
@@ -217,8 +217,9 @@ static int __init vsyscall_init(void)
 	BUG_ON((unsigned long) &vtime != VSYSCALL_ADDR(__NR_vtime));
 	BUG_ON((VSYSCALL_ADDR(0) != __fix_to_virt(VSYSCALL_FIRST_PAGE)));
 	map_vsyscall();
-	sysctl_vsyscall = 1;
+#ifdef CONFIG_SYSCTL
 	register_sysctl_table(kernel_root_table2, 0);
+#endif
 	return 0;
 }
 
diff --git a/arch/x86_64/kernel/x8664_ksyms.c b/arch/x86_64/kernel/x8664_ksyms.c
index a43dedb58fa..68ec03070e5 100644
--- a/arch/x86_64/kernel/x8664_ksyms.c
+++ b/arch/x86_64/kernel/x8664_ksyms.c
@@ -193,8 +193,9 @@ EXPORT_SYMBOL(smp_num_siblings);
 extern void do_softirq_thunk(void);
 EXPORT_SYMBOL(do_softirq_thunk);
 
-void out_of_line_bug(void);
+#ifdef CONFIG_BUG
 EXPORT_SYMBOL(out_of_line_bug);
+#endif
 
 EXPORT_SYMBOL(init_level4_pgt);
 
diff --git a/arch/x86_64/mm/fault.c b/arch/x86_64/mm/fault.c
index e0330921676..5d6b2114cc9 100644
--- a/arch/x86_64/mm/fault.c
+++ b/arch/x86_64/mm/fault.c
@@ -234,6 +234,8 @@ static noinline void pgtable_bad(unsigned long address, struct pt_regs *regs,
 
 /*
  * Handle a fault on the vmalloc or module mapping area
+ *
+ * This assumes no large pages in there.
  */
 static int vmalloc_fault(unsigned long address)
 {
@@ -272,7 +274,10 @@ static int vmalloc_fault(unsigned long address)
 	if (!pte_present(*pte_ref))
 		return -1;
 	pte = pte_offset_kernel(pmd, address);
-	if (!pte_present(*pte) || pte_page(*pte) != pte_page(*pte_ref))
+	/* Don't use pte_page here, because the mappings can point
+	   outside mem_map, and the NUMA hash lookup cannot handle
+	   that. */
+	if (!pte_present(*pte) || pte_pfn(*pte) != pte_pfn(*pte_ref))
 		BUG();
 	__flush_tlb_all();
 	return 0;
@@ -346,7 +351,9 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	 * protection error (error_code & 1) == 0.
 	 */
 	if (unlikely(address >= TASK_SIZE)) {
-		if (!(error_code & 5)) {
+		if (!(error_code & 5) &&
+		      ((address >= VMALLOC_START && address < VMALLOC_END) ||
+		       (address >= MODULES_VADDR && address < MODULES_END))) {
 			if (vmalloc_fault(address) < 0)
 				goto bad_area_nosemaphore;
 			return;
diff --git a/arch/x86_64/mm/ioremap.c b/arch/x86_64/mm/ioremap.c
index 74ec8554b19..58aac23760e 100644
--- a/arch/x86_64/mm/ioremap.c
+++ b/arch/x86_64/mm/ioremap.c
@@ -133,7 +133,7 @@ ioremap_change_attr(unsigned long phys_addr, unsigned long size,
 					unsigned long flags)
 {
 	int err = 0;
-	if (flags && phys_addr + size - 1 < (end_pfn_map << PAGE_SHIFT)) {
+	if (phys_addr + size - 1 < (end_pfn_map << PAGE_SHIFT)) {
 		unsigned long npages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
 		unsigned long vaddr = (unsigned long) __va(phys_addr);
 
@@ -214,7 +214,7 @@ void __iomem * __ioremap(unsigned long phys_addr, unsigned long size, unsigned l
 		remove_vm_area((void *)(PAGE_MASK & (unsigned long) addr));
 		return NULL;
 	}
-	if (ioremap_change_attr(phys_addr, size, flags) < 0) {
+	if (flags && ioremap_change_attr(phys_addr, size, flags) < 0) {
 		area->flags &= 0xffffff;
 		vunmap(addr);
 		return NULL;
@@ -251,7 +251,7 @@ void __iomem *ioremap_nocache (unsigned long phys_addr, unsigned long size)
 
 void iounmap(volatile void __iomem *addr)
 {
-	struct vm_struct *p, **pprev;
+	struct vm_struct *p;
 
 	if (addr <= high_memory) 
 		return; 
@@ -260,24 +260,11 @@ void iounmap(volatile void __iomem *addr)
 		return;
 
 	write_lock(&vmlist_lock);
-	for (p = vmlist, pprev = &vmlist; p != NULL; pprev = &p->next, p = *pprev)
-		if (p->addr == (void *)(PAGE_MASK & (unsigned long)addr))
-			break;
-	if (!p) { 
-		printk("__iounmap: bad address %p\n", addr);
-		goto out_unlock;
-	}
-	*pprev = p->next;
-	unmap_vm_area(p);
-	if ((p->flags >> 20) &&
-		p->phys_addr + p->size - 1 < virt_to_phys(high_memory)) {
-		/* p->size includes the guard page, but cpa doesn't like that */
-		change_page_attr(virt_to_page(__va(p->phys_addr)),
-				 p->size >> PAGE_SHIFT,
-				 PAGE_KERNEL);
-		global_flush_tlb();
-	} 
-out_unlock:
+	p = __remove_vm_area((void *)((unsigned long)addr & PAGE_MASK));
+	if (!p)
+		printk("iounmap: bad address %p\n", addr);
+	else if (p->flags >> 20)
+		ioremap_change_attr(p->phys_addr, p->size, 0);
 	write_unlock(&vmlist_lock);
 	kfree(p); 
 }